- Email: [email protected]
Fetching better deals from creditors: Board busyness, agency relationships and the bank cost of debt
Journal Pre-proof Fetching better deals from creditors: Board busyness, agency relationships and the bank cost of debt
Vu Quang Trinh, Abdullah Aljughaiman, Ngan Duong Cao PII:
S1057-5219(20)30116-2
DOI:
https://doi.org/10.1016/j.irfa.2020.101472
Reference:
FINANA 101472
To appear in:
International Review of Financial Analysis
Received date:
3 September 2019
Revised date:
24 February 2020
Accepted date:
26 February 2020
Please cite this article as: V.Q. Trinh, A. Aljughaiman and N.D. Cao, Fetching better deals from creditors: Board busyness, agency relationships and the bank cost of debt, International Review of Financial Analysis(2020), https://doi.org/10.1016/ j.irfa.2020.101472
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier.
Journal Pre-proof
Fetching Better Deals from Creditors: Board Busyness, Agency Relationships and the Bank Cost of Debt Vu Quang Trinh*, Newcastle University Business School, Newcastle University, United Kingdom
-p
ro
of
*Corresponding author. Email: [email protected] Postal address: Newcastle University Business School, 05 Barrack Road, Newcastle Upon Tyne, United Kingdom. Postcode: NE1 4SE
Abdullah Aljughaiman
re
School of Business, King Faisal University, Saudi Arabia
lP
Email: [email protected] Postal address: Al-Ahsaa 31982, KSA
na
Ngan Duong Cao
University of Bath School of Management, United Kingdom
Jo ur
Email: [email protected] Postal address: Claverton road, Bath, BA 2 7AY, UK.
Journal Pre-proof
ABSTRACT In a cross-country setting, we document that busy boards of directors (i.e., outside directors with multiple directorships) enhance a bank’s financing capacity by lowering its cost of debt, which is consistent with the signalling quality hypothesis. Our analysis further reveals that this negative association is more pronounced in conventional banks than their Islamic counterparts. Possibly owning to the distinctive governance structure and the complexity of the Islamic business model, which requires closer monitoring, Muslim debtholders might depreciate a busy board of directors as it is likely to associate with lower scrutinising effectiveness. Our results provide a positive counterpoint to the negative relationship that exists between busy directors and firm performance, and contributes to understanding the indispensable role busy boards play in debt financing.
-p
ro
of
Keywords: Board busyness, Islamic banks, Conventional banks, Cost of debt JEL classification: C23 G01 G21 G28 L50 M4
Jo ur
na
lP
re
Declarations of interest: none
Journal Pre-proof 1. Introduction “Given that bank loans are the primary source of financing for most economies, the cost of debt plays a significant role in an economy’s growth and performance” (Chen et al., 2016, p.70) Debt contracts are reflected as a central mechanism to economic growth because they provide a useful source to finance investments (Daher, 2017). Like other industrial firms, outside equity and debt remain the two primary finance sources of banks’ capital. Nevertheless, relatively to the former source, the debt market appears to be a predominant source of financing of banks, which is evident through its tremendous size1 and its large share
of
in a typical bank’s total assets. Hence, banks are well-acknowledged as relatively highly
ro
leveraged and heavily regulated as compared to non-financial firms. Given the importance of debt finance to the banking industry, it is worthy of special attention from academic
-p
researchers and practitioners.
Conventionally adhered to the agency theory, interest conflicts among the three parties
re
comprising of directors, shareholders, and debtholders are inevitable (Chakravarty and
lP
Rutherford, 2017). Particularly, shareholders of high-leveraged institutions, including banks often have a strong incentive to expropriate wealth from debtholders through investing in
na
risky projects (Fields et al., 2012). Debtholders may anticipate these incentives, and thus, require greater risk premium (returns) on their capital, leading to a higher cost of debt (see Jensen and Meckling, 1976; Ertugrul and Hegde, 2008). Creating a high-quality board of
Jo ur
directors (BOD hereafter) to scrutinise bank activities and investments on behalf of debtholders is a fundamental solution for such conflict of interest. Numerous existing studies (e.g., Anderson et al., 2004; Fields et al., 2012) have focused on the linkage between quantifiable measures of the management quality (i.e., corporate governance and board attributes) and various corporate outcomes such as cost of debt, equity return, and firm value. Some of them demonstrate the effect of the BOD characteristics on the cost of debt financing in non-financial firms (e.g., Anderson et al., 2004; Piot and MissonierPiera, 2007; Lorca et al., 2011). They generally claim that debtholders favour strong monitoring mechanisms, which tend to limit managerial opportunism. Thus, the quality of governance mechanisms can be considered as a channel, which may reduce creditors’ risk, and therefore, debtholders might become more lenient in the way the risk premium is set. In 1
Statistics of Bank for International Settlements (2017) show that at the end of the first quarter of 2017, the global debt securities outstanding reached to 21,749 billion of US dollars; among this, international debt securities outstanding of banks (6,427 billion of US dollars) are double than that of non-financial firms (3,261 billion of US dollars).
1
Journal Pre-proof other words, the increased monitoring effectiveness of the board towards the management team and CEO as a default risk-reduction corporate governance mechanism may benefit the debtholders as it aligns the interest conflicts between them and agents. This raises the high probability that the bank can meet the interest of the creditors and principal payment obligations. Although those studies help understand the impacts of management quality on different corporate outcomes, the research is still limited within the contexts of busy boards and cost of debts. As delegated monitors, outside directors are charged with legal roles and accountabilities offering effective oversight over executives or managers such that conflicts
of
of interest between creditors, shareholders and managers are minimised. Debtholders,
ro
therefore, can benefit from outside directors’ effective scrutinising through mitigation in agency problems such as shirking, perquisite consumption, and overinvestment (Trinh et al.,
-p
2019). Such decreased agency cost, together with the reduction in information asymmetry, can mitigate the cost of debt financing (Chen and King, 2014).
re
The existing literature studying cost of debt usually excludes banks from their samples, except for Penas and Unal (2004), Deng et al. (2007), Deng et al. (2017). However, those
lP
banking studies concentrate solely on the association between the borrowing rate and nongovernance aspects such as diversification (i.e., geographic, assets and non-traditional
na
activities) (e.g., Deng et al., 2007; Deng et al., 2017). Therefore, based on our review of literature, the association between the BOD busyness and cost of debt remains undiscovered,
Jo ur
specifically for financial firms.
As discussed earlier, debt financing is considered as the main source of funds in the banking industry. Thus, banks need to assure the achievement of an effective management to their debt terms. A highly qualified board in the banks can achieve this important goal. Fields et al. (2012) argue that the quality of the board might contribute to influence on the cost of debt capital. Taking in consideration the significant influence of busy board on the board quality, it is important to investigage the busy board in the context of banking industry. A board that is characterised as busy may affect the board quality positively through their experience and knowledge. This may lead to greater monitoring and better quality advice to the management, which may mitigate the banks cost of debt. Accordingly, our study extends the cross-country banking literature by answering the following important question: Does busy BOD reduce the cost of debt in the banking sector? The study brings critical values since debt financing is one of the crucial capital sources for bank operations (leveraged), and we are one of the first to look at this specific research 2
Journal Pre-proof question. While many prior studies show a negative effect of board busyness and firm performance, which is considered as problematic for the companies’ shareholders (e.g., Fich and Shivdasani, 2006; Cashman et al., 2012), there is very limited evidence showing that whether lenders prefer a firm’s board busyness or not (except for Chakravarty and Rutherford, 2017). Furthermore, since our sample covers countries that use dual banking operating systems (Conventional and Islamic), special attention has been paid to Islamic banks. Our focus on these two bank types is crucial under the on-going debate about the resilience and stability of the Islamic banks. Some researchers cast doubt on the long-run performance and risk-taking
of
of Islamic banks and underpin the need for further investigations for their governance
ro
structure and board efficacy (e.g., Cihák and Hesse, 2010; Beck et al., 2013; Abdelsalam et al., 2016). While conventional banks operate within the traditional interest-based framework,
-p
their Islamic counterparts refer to a banking system that operates under risk-sharing base. Specifically, Islamic banks must follow several Shari’ah principles that distinct their
re
operating system and have its own multi-layer governance system (Mollah and Zaman, 2015)2. BODs and managers in Islamic banks are required to adhere to Islamic principles of
lP
Shari’ah in fulfilling their missions to maximise shareholders’ wealth. The stakeholders’ interests in this bank type might extend beyond the financial interest to involve ethical and
na
religious values/needs (Alnasser and Muhammed, 2012). The failure to meet those expectations by the board and managers creates an extra source of agency problems. Given
Jo ur
these differences between conventional and Islamic banks, this paper is inspired by the demand to recognise the differences in busy board influences between two bank types. We argue that the need for monitoring of a BOD towards managers in Islamic banks tends to be higher than that in their conventional counterparts, leading to dissimilar effects of BOD busyness on cost of debt. We also believe that a comparative investigation of BOD busyness across two different bank types, Islamic and conventional banks, is particularly important to on-going debate on influences of board busyness on financial indicators when levels of agency and monitoring demands are different. This, in turn, can improve our understanding of busy directors’ role in unique bank governance. Therefore, we next aim to investigate whether the distinct features of Islamic banks alter the relationship between the cost of debt and busy board. 2
To be compliant with Sharia’ah Principles, Islamic banks are not allowed to charge interest on money, take excessive risks, invest in harmful activities (e.g., gambling, alcohol,), and their transaction must be based on real economic assets. Islamic banks financial instruments must be based on profit-losses sharing contracts.
3
Journal Pre-proof Our second research question is set upon on the argument that busy directors appear to have differential effects on bank performance depending on the level of agency conflicts related to the corporate operations (Trinh et al., 2019). Considering the monitoring and advising roles of directors, the value of holding multiple directorships depends ultimately on the relative importance of effective oversight, and the governance attributes. If the need for scrutinising is vital for certain types of firms, because of the greater business-process complexities and uniqueness of agency problems, the monitoring role of independent directors is relatively more important than their resource function. To investigate the distinct effects of board busyness on the cost of debt conditional on banks’ institutional
of
characteristics and business models, we evaluate such issue across two global banking
ro
systems: Islamic and conventional banks.
The data focuses on publicly traded commercial banks over the estimated period of 2010-
-p
2015. Several interesting results are obtained. First, we find a negative association between the banks’ cost of public debt and a busy BOD on our full sample. This confirms our main
re
hypothesis that busy BOD significantly reduces the cost of debt in banking sector. Second, the opposite signs of coefficients of interaction terms between busy boards and Islamic bank
lP
dummy factor suggest that the negative relationship between the cost of debt and a busy board is lessened for Islamic banking. In other words, by conditional on the bank types, we
na
find that Islamic banks having busy outside directors are less likely to enjoy lower cost of debt comparing to their conventional counterparts. However, we find a non-linear
Jo ur
relationship between busy BOD and cost of debt for conventional bank subsample, which is not significant in Islamic bank subsample. Our results are robust across different board busyness measures and various model specifications. Our findings contribute to the line of research that focuses on the relationship between board attributes and the cost of debt (e.g., Lorca et al., 2011; Chakravarty and Rutherford, 2017); and also complement comparative governance literature underpinning the Islamic conventional banking paradigm (e.g., Mollah and Zaman, 2015; Mollah et al., 20173; Trinh et al., 2019) by investigating the impact of the board busyness on a bank’s borrowing rates from an international perspective. For previous studies, while Mollah and Zaman (2015) examine a specific character of Shari’ah board (i.e. the board size) and its influence on bank performance, Mollah et al. (2017) continue investigating the relation between general
3
Mollah and Zaman (2015) examines a specific character of SSB (i.e. the board size of SSB) and its influence on bank performance. Mollah et al. (2017), investigates the relation between general corporate governance structures, via a developed index, and bank risk-taking across the two bank types.
4
Journal Pre-proof corporate governance structures, via a developed index, and bank risk-taking across the two bank types. More recently, Trinh et al. (2019) and Elnahass et al. (2019) have contributed to Mollah papers by testing the effect of board busyness (both BOD and Shari’ah board) on bank performance, risk-taking behaviour and market valuations. However, all these studies consider general corporate governance and/or busyness effects under the views of investors and banks. Our paper has therefore incrementally contributed to existing corporate governance research in Islamic and conventional banking system by comparatively evaluating board busyness function in a linkage with lenders or debtholders. This ends up with a consideration of bank cost of debt and board busyness within the specific context of
of
those both bank types.
ro
Furthermore, although Islamic banks are prohibited from paying an interest rate to their debtholders, they still pay an expected profit rate comparable to interest rates on other
-p
savings accounts. In this study, we also contribute to the method of measuring the cost of debt. While most of the existing studies on the cost of debt mostly focused on interest rates
re
when measuring a firm’s cost of debt (e.g. Reeb et al., 2001; Klock et al., 2005; Cremers et al., 2007; Kisgen and Strahan, 2010; Chen and King, 2014), this research uses a new
lP
alternative metrics: after-tax weighted average cost of debt (ATWACD). We argue that a bank often uses various sources, including either short- or long-term debt, to finance their
na
activities; therefore, using only interest rates as the cost of debt appears to fail in reflecting the lending price of a bank fully and accurately. ATWACD hence is more appropriate
Jo ur
because of its strong representation in either long-term or short-term debt. More importantly, our new measure is particularly useful in estimating the relative cost of debt of Islamic banks, which are known as interest-free businesses. Through this proxy for the cost of debt, analysis can reasonably evaluate whether and how debtholders value the busy boards of directors, especially those in the interest-free financial system like Islamic banks. The next section reviews the literature on the motivations for busy directors and the link between them and the cost of debt; and then sets the tested hypotheses. Section 3 describes our sample and database sources, and the methodology. Section 4 reports the descriptive statistics. Section 5 presents the empirical results, which is followed by additional testing and robustness checks (Section 6). Section 7 presents a conclusion. 2. Research background and empirical predictions 2.1. Sources of busy directors’ benefit: reduction in the cost of debt
5
Journal Pre-proof An effective bank BOD is one of the crucial components of strong firm governance, especially after recent financial turmoil. Nevertheless, the governing bodies of the BOD appear to provide limited regulation regarding specific guidance on how BOD members are anticipated to execute their role. Furthermore, the BOD functioning (e.g. BOD size, independence, directorships) is in general open to interpretation. For instance, the common question that if larger (or smaller), diverse (or uniform), and highly independent (or low independent) BOD within banks is more effective are still vague (e.g., Elyasiani and Zhang, 2015). Similarly, findings from present works are also equivocal with regards to the true nature of the actual role of busy directors (those are sitting in several for-profit firms’ boards)
of
on bank financial decisions (Elnahass et al., 2019). Advocates support that the rich sources
ro
brought by busy directors are beneficial to banks whilst opponents provide detrimental evidence.
-p
Corporate finance theorists assert that the value of a firm/bank is the present value of future cash flows (CFs) (Chen and King, 2014). Thus, the effects of busy BOD on bank value
re
can be from at least two sources: the impact on the CFs stream and/or the effect on the cost of capital that future CFs are discounted. As highlighted by Easley and O’Hara (2004), the
lP
fundamental in firm policies is the cost of capital due to its influence on profitability, and thereby investment decisions. Chakravarty and Rutherford (2017) is a notable recent work
na
providing empirical evidence for non-financial firms in a single country (US) through a hostile takeover framework that board busyness contributes to the reduced cost of debt.
Jo ur
However, the association of busy boards and the cost of debt, particularly in the banking industry, within previous research is still unanswered. Our study is different to Chakravarty and Rutherford by three-fold. First, we conduct a study on a global basis while their study is a single-country research. Second, we explore BOD busyness-cost of debt nexus in the context of banking. It is important to investigate such association in this sector as banks characteristics and business models are different from non-financial firms. Particularly, banks have high leverage level, encounter contagion risk in the industry, work in a significantly regulated environment, and have important impacts on the real economy (Elyasiani and Zhang, 2015). Furthermore, the banking industry is subject to unique agency costs that are different from the conventional agency problems (Safieddine, 2009). These different agency costs might distinguish the influence of busy board on the cost of debt in the banking industry. Consequently, we have addressed the lacking attention paid to the busy board of banks. Third, our study takes a step further to conduct a comparison research between two alternative banking models: conventional and Islamic banks. 6
Journal Pre-proof Existing corporate theories advise a close and important BOD busyness-borrowing costs nexus. In general, the theoretical foundation for the association between busy board and cost of debt could be driven from the managerial power theory (Bebchuk et al., 2002; Bebchuk and Fried, 2004). This is consistent with Pathan et al. (2019), who discuss that this theory recognises the important monitoring role that provided by the independent director. According to the managerial power theory, a firm with poor monitoring allow managers to acquire stronger power to benefit their interests that is relative related to the firm performance (Bebchuk et al., 2002; Bebchuk and Fried, 2004). This might give incentive to managers to take higher risk level, which maximise the shareholder wealth, on the expense of debt
of
holders. In addition, Fama (1980), Fama and Jensen (1983) argue that a busy board, that have
ro
a number of director memberships, can be an indication of his reputation as an effective monitoring of corporate managers. Taking together, the managerial power theory with this
-p
positive view, we can theoretically expect a negative relationship between busy board and cost of debt.
re
In term of empirical research, we thoroughly discuss below three main channels for how busy BODs could affect the bank cost of debt and develop the hypotheses accordingly.
lP
Our first arguable channel is related to weaker shareholder control. Weaker shareholders control could be created by having busy outside directors. The rationale is that directors held
na
additional directorship appointments tend to weaken shareholder control as they pursue conservation of their directorship reputation through a possible channel: putting anti-takeover
Jo ur
provisions in place (Chakravarty and Rutherford, 2017). A study by Cremers et al. (2007) suggests that not all board characteristics have the same influence on debtholders as they have on shareholders. For instance, shareholders prefer a strong shareholder control to maximise their wealth, while weak shareholder control is strongly desired by the debtholders to minimise the asset substitution problem. Thus, given such shareholder control-weakening behaviour of busy outside directors’ appeals to creditors, the benefits brought by those busy BODs are expected to outweigh the distraction risk to the bank stability (e.g. Faleye et al., 2011). This makes the lenders perceive busy boards positively, which may signal their preference by providing loans to firms at preferential rates (e.g. Shivdasani, 1993; Klock et al., 2005; Cremers et al., 2007). This supports a positive view that the number of other jobs holds by a “busy” directors may serve as a measure of his/her reputation as an effective monitor of corporate managers (Fama, 1980; Fama and Jensen, 1983), which diminishes managerial opportunism and mismanagement and hence, lower the cost of debt capital (Reeb et al., 2001; Ertugrul and Hegde, 2008). Furthermore, companies which protect their outside 7
Journal Pre-proof directors from a potential lawsuit, and hence allow these directors to pursue their self-interest at the expenses of shareholders, might enjoy a lower cost of debt (Bradley and Chen, 2011). Bradley and Chen (2011) also ascribe the association between the firm cost of debt and busy directors to the weaker shareholder control mechanism, because weaker shareholder control mitigates the asset substitution risk. Another argument that describes the association between the busy board and the cost of debt is through the insolvency risk channel. Trinh et al. (2019) find an inverse relationship between insolvency risks and busy outside directors within banks. Specifically, the probability of insolvency is considerably greater when BODs fail to fulfil their legal
of
responsibilities. Because multiple directorships help enhance outside directors’ ability,
ro
experience and networking to provide better advising and monitoring services for banks, insolvency risk is reduced. In particular, Elyasiani and Zhang (2015) find that busyness of
-p
directors improves their performance and consequently lowers the expected cost of financial distress. Wider networking of busy BOD directors helps banks to easier find good sources for
re
financing investments, leading to a reduction in the cost of borrowing. We hypothesise that busy BODs should be associated with a lower cost of debt. We further argue that the
lP
busyness of directors leads to a higher reduction on the cost of debt for banks with a higher insolvency risk. Banks with higher bankruptcy risk-benefit most from busy outside directors
na
due to their higher probability and costs of financial distress. In contrast, banks with little bankruptcy risk-benefit least from busy directors due to their unlikely default.
Jo ur
In a similar vein, companies with risky debt tend to forgo positive net present value investments if most value of the investment does toward the debtholders as poor states arise (Myers, 1977). Busy BODs can enhance the bank board advising and monitoring effectiveness through a decline of the probability of the poor states arising. Thereby, owners may have higher incentives as well as the confidence to invest their capital into potential value-enhancing projects (Elyasiani and Zhang, 2015). In other words, BOD busyness may lead the board following the optimal investment policy through making sufficient internal resources and enjoying a low cost of capital and hence, curtails the underinvestment problem when a bank has growth opportunities and a greater cost of external financing. In our particular case, banks having busy BODs may have a lower probability of the poor states occurring, and consequently, they should be the preference of the creditors. Based on the above, we reasonably contend that busy BODs should lead to a lower cost of debt by increasing the board effectiveness and mitigating the underinvestment problems.
8
Journal Pre-proof To sum up, our previous arguments indicate that effective BOD (e.g., busy board) can achieve high quality of management practices. Intuition contends that higher quality of management practices might fetch better deals from debtholders. Because BOD is a crucial input in the bank operating process, higher quality of BOD (better input) might lead to more efficient operations and thereby lower the bank underlying business risk. This, consequently, might relax the terms of credit the banks could get from their debtholders (Rahaman and Zaman, 2013). The simple economic framework adopted by Rahaman and Zaman (2013) can illustrate why the BOD management quality may substantially influence a bank’s borrowing cost via three points. First, although debtholders cannot observe the intrinsic quality of a
of
bank, the high-quality bank could always formulate costly signalling mechanisms/devices to
ro
fetch better deals from debtholders. The signal can be motivated in terms of reputational capital (Diamond, 1989) or relationship lending (Boot and Thakor, 1994) or by adopting
-p
better quality of management practices (Bloom and Van Reenen, 2007). All these motivations could be obtained from the multiple directorships of the directors. Second, debtholders may
re
not have sufficient motivation to fully scrutinise the bank-quality spectrum because large debtholders could always diversify away borrower-specific risks in their loan portfolios.
lP
Therefore, signalling by banks rather than monitoring by debtholders could be a better channel to understand the interaction between board quality and debt contract terms. Third,
na
the benefit from signals (adopting costly managerial practices) may not be sufficiently high for debtholders although those management quality practices enhance the firm total value
Jo ur
(Bloom and Van Reenen, 2007). It is therefore important to note that as long as signals from busy BODs are costly and benefits from signalling are sufficiently high, we should observe heterogeneity in BOD busyness across banks, and the variation in multiple directorships of BODs should also be correlated with the heterogeneity in the bank’s cost of debt. This argument is in line with Bloom and Van Reenen (2007) and Bloom et al. (2011) who provide agency explanations of why poorly-managed firms exist. Taken together, we hypothesise that board busyness is likely to reduce the bank cost of debt. Accordingly, we set our first hypothesis in the alternative forms: H1: Banks with busy boards exhibits lower cost of debt than banks with non-busy boards 2.2. Director busyness and the cost of debt: the matter of bank type Islamic banks differ primarily from conventional banks regarding the additional layer of governance of the former, which aims to overlook banks’ religious affairs. Particularly, the latter has no concerns on any religious matters whilst the former are required to conduct its 9
Journal Pre-proof activities in compliance with Islamic laws. In details, Islamic banks have to follow Shari’ah principles, which (1) prohibits payments or receive of interest on the debts (Riba), (2) relay on risk-sharing bases instead of risk-shifting for all financial transactions, (3) prohibits excessive risks and a certain type of investment activities (e.g., Alcoholic). To comply with these principles, the nature of Islamic banks’ liability side (e.g., financial contracts) has to be different from their counterparties (conventional banks). Even though interest in lending is prohibited in Islam, some of their transaction could be categorized as debt-based financing (Abedifar et al., 2013). In these financing contracts, the financer buys or ask to construct the underlying assets then sell it to the client using the deferred-payment basis with one or
of
several instalments. Another financing source that Islamic banks rely on is investment or
ro
saving accounts, which generates a return based on agreed profit rates. Abedifar et al. (2013) argue that the payoff to investment account holders in Islamic banks
-p
dependents on the bank’s performance and depositor’s religiosity. Religious depositors might accept lower return (cost of debt for Islamic banks) if the bank assures their compliance with
re
Shari’ah principles even if the bank's performance is poor. Therefore, their depositors might care more about the ethical and religious issues in the banks. In contrast, religious depositors
lP
might be risk-averse and demanding higher return if the bank showing a risky behaviour (Abedifar et al., 2013). Thus, the Islamic banks’ cost of debt can be affected negatively.
na
Given the unique characterises of Islamic banks and the additional restrictions in their business models, their BODs might encounter complicated and more responsibilities that
Jo ur
need to be achieved (e.g., compliant with Shari’ah principles). Aljughaiman and Salama (2019) discuss that the BOD responsibilities to comply with Shari’ah principles might lead to more constraints on their ability to monitor and advice management and manage risk, which might lead to differential impacts on their risk-taking behaviour. Thus, the differences in agency relations between Islamic and conventional banks can produce a different relationship between busy boards and cost of debt. Furthermore, board members are not Shari’ah scholars, and thus not able to confirm the bank's compliance with Shari’ah principles. Therefore, Islamic banks’ debtholders may not appreciate the busy board, and thus not have a strong influence on the cost of the fund they provide to the bank. In this study, we hypothesise that conventional banks employing busy BOD may enjoy lower cost of debt than Islamic banks with busy BOD. This is well explained by the finding of Trinh et al. (2019) which find that busy BODs in Islamic banks show detrimental effects on bank financial stability while those in conventional banks still bring reputational benefits
10
Journal Pre-proof such as networking and experience to the firms. This leads us to the second hypothesis as below: H2: Conventional Banks with busy boards exhibit lower cost of debt than Islamic banks with busy boards 3. Data and Research method 3.1. Data collection and sample To test our hypotheses, we employ several sources including the Thomson One Reuters DataStream, Bankscope, annual bank reports, and World Bank database.4 We focus on a
of
mixed sample of Islamic and conventional banks listed on stock markets from 2010 to 2015. This estimated period allows us to avoid the potential effect of the crisis shock of 2007-2009
ro
that banks encountered. We started with an initial sample of 3038 banks (196 Islamic banks and 2842 conventional banks) operating cross 36 countries. We then dropped any bank that
-p
have headquartered in the countries having only Islamic or conventional banks and having
re
less than two listed banks; dropped any other banks which do not have official websites and accounting period from 01 of January to 31 of December; excluded full investment banks and
lP
conventional banks with Islamic window (i.e., conventional banks with separate Islamic banking department with a supervision of Shari’ah board); and finally we do not include
na
banks which have less than three consecutive year’s full data availability. Our criteria are in line with the study of Elnahass et al. (2019) and other previous banking literature such as
Jo ur
Mollah et al. (2017).
Our ultimate dataset includes 386 firm-year observations (70 publicly quoted commercial banks operating in 11 countries). We report the sample distribution by country and bank in Table 1: 43 conventional banks (236 observations) and 27 Islamic banks (150 observations). We assume that the differences in Islamic banking models across countries due to various country-specific regulations do not exist. [Insert Table 1 here] 3.2. Research method
4
Consolidated bank accounting and market data were drawn from Datastream and Bankscope database. We manually collect corporate governance data (i.e. information of independent non-executive directors and the board of directors) from annual reports. We define a director as busy if he/she holds at least three directorships or two outside/additional directorships. We count the number of directorships held by directors in all for-profit private and public firms, consistent with previous studies (Elyasiani and Zhang, 2015; Chakravarty and Rutherford, 2017; Trinh et al., 2019). And finally, we collect macroeconomics indicators from the World Bank database.
11
Journal Pre-proof 3.2.1 Measuring agency cost of debt The primary dependent variable of our analysis is the bank cost of debt. We measure this by the after-tax weighted average cost of debt for the security, which is available in Bloomberg database. It is calculated by using government bond rates, a debt adjustment factor (DAF), the proportions of short-term and long-term debt to total debt, and the stock's effective tax rate. The debt adjustment factor represents the average yield above government bonds for a given rating class. The lower the rating, the higher the adjustment factor. The DAF is only employed when a bank does not have a fair market curve. When a bank does not have a credit rating, an assumed rate of 1.38 (the equivalent rate of a BBB+ Standard &
of
Poor's long-term currency issuer rating) is employed. The exact calculation of DAF is as
ro
follows:
-p
Cost of Debt = [[(SD/TD) * (CS * AF)] + [(LD/TD) * (CL * AF)]] * [1-TR] Where: SD is Short-term Debt. TD is total debt. CS is Pre-Tax Cost of Short-term Debt.
re
AF is Debt Adjustment Factor. LD is Long-term Debt. CL is Pre-Tax Cost of Long-term Debt.5 TR is Effective Tax Rate. Our measure is useful in estimating the relative cost of debt
lP
of Islamic banks, which are known as interest-free businesses. Accordingly, traditional interest rates are replaced by an equation, including government bond rates (short-term and
na
long-term debt) that both banking models can employ to finance for their operations.
Jo ur
3.2.2. Measuring the busyness of boards of directors A busy director is defined as an individual who serves in at least two outside firms. This definition is similarly employed in previous banking studies such as Elyasiani and Zhang (2015) and Trinh et al. (2019). Based on this, two measures of busy board of directors (BBOD) are used. The first measure is the ratio of outside directorships per outside director (ABOD), representing the average number of other outside board seats held by each outside director. It is computed as the total number of additional (outside) boards occupied by outside directors divided by the number of outside directors on the board (Ferris et al., 2003). The second measure is the percentage of busy outside directors (%BBOD), which is calculated as
5
In the case of Islamic bank, the cost of long-term debt is the return rate that the investment account holders receive from the bank.
12
Journal Pre-proof the number of outside directors serving on two or more additional (outside) firms divided by the number of outside directors on the board.6 3.2.3. Empirical models To examine our hypotheses, we employ the traditional ordinary least square (OLS) with robust standard error approach. Accordingly, we present the main equations model as follows: 𝐶𝑜𝑠𝑡 𝑜𝑓 𝐷𝑒𝑏𝑡𝑖𝑡 = 𝛽0 + 𝛽1 𝐵𝐵𝑂𝐷𝑖𝑡 + 𝛽2 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 + 𝛽3 𝐵𝐵𝑂𝐷𝑖𝑡 ∗ 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 + 𝜙𝑃 + µ𝑌𝑒𝑎𝑟 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝜀𝑖𝑡 𝐶𝑜𝑠𝑡 𝑜𝑓 𝐷𝑒𝑏𝑡𝑖𝑡 represents
the
bank
cost
of
Where,
(1) of
debt;
𝐵𝐵𝑂𝐷𝑖𝑡
represents
ro
{ABOD; %BBOD}; 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 represents the Islamic bank dummy variable, taking a value of
-p
1 if the observed bank is classified as Islamic, otherwise 0; 𝐵𝐵𝑂𝐷𝑖𝑡 ∗ 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 is the interaction term between busy BOD and ISLAMIC dummy factor, which is used to control
re
for the mediating effects of the two different bank types; 𝜙𝑃 is a vector of control variables in the cost of debt model.
lP
Following prior studies (e.g., Lorca et al., 2011; Fields et al., 2012; Chakravarty and Rutherford, 2017), we control for board as well as firm-level and country-level characteristics
na
in our multivariate regressions to limit potential omitted variables bias. Those are widely employed in the previous cost of debt literature, which captures factors affecting the bank
Jo ur
borrowing rate. We first include a set of bank governance characteristics such as board size (LogBSIZE) and board independence (%INDEP) and CEO duality (DUAL). The former is measured by the number of directors on the board while the latter is measured by the percentage of outside non-executive directors on the board. The final is calculated as the dummy variable, which takes the value of one if the chair and CEO is the same person, otherwise, zero. We next control for bank and country-specific variables consisting of bank size (LogTA), bank leverage (LEV), GDP per capital (GDPCAPITA), Herfindahl-Hirschman index (HHI), and average index of country governance (GOV_COUNTRY). Table 2 contains a complete list of variable names and definitions/calculations. [Insert Table 2 here] 6
Due to limited data for Islamic banking, we are unable to distinguish between other types of directorships or activities of the director. We understand that this may be the limitation of the proxy of busyness, but this is consistent with previous studies in the field such as Elyasiani and Zhang (2015) and Trinh et al. (2019). In addition, in line with their studies, we have included only paid jobs, which affect significantly directors’ time and monitoring ability. In other words, we do not include voluntary jobs.
13
Journal Pre-proof 4. Descriptive statistics Table 3 provides descriptive statistics for the variables in all estimated models, which includes the cost of debt, busy board of directors, governance attributes, and other control firm-specific and country-specific variables. Our sample of 386 firm-years shows that the average and median cost of debt (Cost of Debt) are around 2.281% and 1.394%. Our t-test proves that the cost of debt of conventional banks - CBs (2.878%) is significantly higher than that of Islamic banks - IBs (1.343%). As for the board governance structure and board busyness, we find that, on average, the board of directors for the whole sample (IBs, CBs) composed of slightly fewer than 10
of
directors (11, 9) (BSIZE), which is consistent with the average number of directors on board
ro
in the study of Chakravarty and Rutherford (2017). An average director in our estimated sample holds about two outside directorship seats (ABOD), which is also similar to other
-p
previous research (Fich and Shivdasani, 2006; Cashman et al., 2012; Elyasiani and Zhang, 2015). This number is higher in CBs (2.374) compared to IBs (2.071); however, the t-test
re
indicates an insignificant difference between the two subsamples.
lP
Busy directors make up about 47.9% of a board of directors (%BBOD), 43.4% of an IB board and 50.85% of a CB board. We find that the percentage of busy directors on board in CBs is higher than that in IBs, showing that board of directors in listed CBs appears to be
na
busier than the board in listed IBs, supported by the result of two-sample t-test. In addition, we find that the percentage of independent directors (%INDEP) on the board for the full
Jo ur
sample (IBs, CBs) averagely is 35% (33%, 36%). [Insert Table 3 here]
Table 4 shows the results of the Pearson pair-wise correlation matrix for the main independent variables in our sample. We find that correlations among explanatory variables are within acceptable limits and raise no serious concerns on multicollinearity. [Insert Table 4 here] 5. Empirical findings 5.1. Busy board of directors and the cost of debt In Table 5 (Model 1 and 2), we show how the bank cost of debt is impacted when a bank adds an additional busy outside director to the board. We find that busy board of directors tends to reduce the bank cost of debt across two alternative proxy measures of board
14
Journal Pre-proof busyness (ABOD and %BBOD), after controlling for other board composition, firm-level and country-level characteristics. This finding is consistent with Chakravarty and Rutherford (2017), in their assertion that lenders recognise banks with busy independent directors as safer investments and are therefore willing to lend their capital at a lower interest rate. Additional studies show the similar concept include Shivdasani (1993), Klock et al. (2005), Cremers et al. (2007), Elyasiani and Zhang (2015), and so on. Economically, the coefficient of average directorships of outside directors suggests that for every additional directorship added to the board, the cost of debt is reduced by about 0.353%. In addition, the cost of debt for banks whose board increases 1% of busy outside
of
directors is reduced by about 1.623%. Based on these results, we conclude that board
ro
busyness has a negative and statistically significant impact on banks’ cost of debt, thus supporting our first hypothesis. Overall findings are consistent with our expectations that
-p
busy BODs tend to bring their reputational benefits to their banks (see Elnahass et al., 2019). This suggests that busy BODs are likely to make their banks more attractive under the eyes of
re
the debtholders, resulting in lower borrowing costs.
We also use interaction terms between busy BODs and IB dummy variable (i.e.
lP
ABOD*ISLAMIC; %BBOD*ISLAMIC) to conduct comparative analyses between IBs and CBs for the impact of busy outside directors on bank cost of debt; our second hypothesis.
na
Results generally show that busy BOD, across two alternative metrics (ABOD, %BBOD), has a significantly higher agency cost of debt in IB system than their conventional counterparts.
Jo ur
This is presented by the highly significant and positive coefficients of interaction terms in both models. This is in line with our expectations and our second hypothesis that CBs having busy BOD is likely to have a lower cost of debt than IBs having busy BOD.7 According to the magnitudes of the estimates of both interaction terms, it is possible for Islamic banks to still benefit from the busy boards for lower cost of debts. However, the reduced levels are lowered by approximately 50% (that results in coefficient equal to -0.195) comparing to those of conventional banks (-0.353) for ABOD. The finding implies that bank debtholders tend to be more concern about the busyness effect (e.g. lax attention, time and efforts) in IBs than that in CBs due to the higher complication of IBs in terms of agency relationships and institutional environments. [Insert Table 5 here] 7
In unreported tests, we run tests for the IB subsample separately by including size of Shari’ah board and multiple directorships of Shari’ah board. The findings are still consistent to our main results.
15
Journal Pre-proof The signs of other control variables are consistent with previous literature. For example, board size (LogBSIZE) is significantly and negatively related to the cost of debt, which suggests that banks having large BOD are likely to reduce the firm cost of debt. This is in line with previous studies of Lorca et al. (2011), Fields et al. (2012) and Chakravarty and Rutherford (2017) showing that larger board is negatively associated with firm borrowing rates. Results of board independence indicate that a high presentation of outside directors in boards is likely to increase the bank cost of debt significantly. This can be explained that banks with higher board independence tend to reduce the firm performance (Yermark, 1996, Wintoki et al., 2012; Pathan and Faff, 2013) and thus, the debtholders may require higher
of
premium from their lending.
ro
5.2. Trend analysis: bank cost of debt by the interval categories of board busyness
-p
We next follow the study of Chakravarty and Rutherford (2017) to provide a trend analysis of a bank average cost of debt for every incremental change in BOD busyness
re
variable. We indeed partition our board busyness proxy into equal increments and investigating how the cost of debt change with each increasing increment category.
lP
Table 6a (Panel A) reveal that we have split %BBOD (i.e. the percentage of busy outside directors serving on the BOD) into 10% segment categories: below 10%, 10-19.99%, 20-
na
29.99%, etc. Results show that at lower board busyness interval (up to 59.99%), the average cost of debt reduces as the percentage of busy outside directors increases. However, an
Jo ur
opposite result is found when the busyness interval exceeds 59.99%. Similarly, Table 6a (Panel B) provide findings on how the bank cost of debt is influenced when a bank adds an additional average outside directorships to the BOD. We explore that with every incremental addition of an average outside directorship to the BOD, the bank cost of debt decreases. However, the cost of debt slightly increases when the average outside directorships exceeds 8.99. These findings are in accord with prior studies in non-financial firms (e.g., Chakravarty and Rutherford, 2017) and in line with our main results. [Insert Table 6a here] 5.3. The possible non-linear effect of busy boards of directors on bank cost of debt We continue testing the possible non-linear effect of busy BOD on bank cost of debt between three subsamples: full sample, CBs and IBs subsample. We find for full sample and
16
Journal Pre-proof CBs subsample that, in Table 6b (Panel A and B, respectively), busy BOD is significantly related to a reduction in bank cost of debt. However, the sign directions of the square of board busyness, i.e. (ABOD)2 and (%BBOD)2 are reversed, implying a non-linear association between busy BOD and the bank cost of debt. This is consistent with the main result in Table 5 and the trend analysis in Section 5.2. We find no significant evidence for the effect of busy BOD on cost of debt in IBs. This finding, again, confirms our second hypothesis that CBs with busy BODs have a lower cost of debt than IBs with busy BODs. [Insert Table 6b here]
of
5.4. Agency relationships, busy board of directors and the cost of debt
ro
Trinh et al. (2019) argue that the agency problem of a bank could be affected by busy BOD through its functions of scrutinising, either positive or negative. We, hence, examine
-p
whether agency costs have any impacts on the main result related to board busyness-cost of debt nexus.
re
In Models 1 and 2 of Table 7, we report the results about the effect of agency costs (i.e.
lP
cash to total assets) on the relationship between busy BOD and bank cost of debt. We find that holding more cash in CBs appears not to show any significant influence on the negative
na
nexus between busy BOD and the borrowing rates. Meanwhile, holding higher cash level in IBs tends to increase such type of costs, which can be explained that debtholders in this bank
Jo ur
type seem not to have trust in the ability of busy outside directors in monitoring the cash use of managers. The findings are consistent with the study of Elnahass et al. (2019) showing that busy BODs in IBs are related to higher agency conflicts. [Insert Table 7 here]
6. Additional Testing and robustness checks 6.1. Effects of busy boards on cost of debt after controlling for additional control variables In this section, we test whether our main results in Table 5 are sensitive to changes in the components of control variables. We add additional independent variables related to corporate governance characteristics comprising of the qualifications of outside directors (i.e. %INDQ), the percentage of outside directors with financial expertise (i.e. %INDEXP). The size of audit committee in natural logarithm form (i.e. LogASIZE) the percentage of busy directors serving in audit committee (i.e. %BAC), and the average number of board meeting
17
Journal Pre-proof held by outside directors (i.e. INDMET). We find in Table 8 that the main findings remain unchanged, implying that our stories are robust when adding more control variables. [Insert Table 8 here] 6.2. Alternative measures of Cost of Debts For more robustness check, we use alternative measures for cost of debts. In details, we use the interest rate on the banks’ debt (i.e., Interest), which is computed as the interest expense for the year divided by the interest-bearing debt (Francis et al., 2005; Bliss and Gul, 2012), as a measure of cost of debt for CBs subsample. However, for the IB subsample, we
of
used the return on IAHs (i.e., Return)8 as an alternative proxy for its cost of debt because IBs
ro
are interest-free businesses. We derived these data from DataStream, Osiris and Bloomberg. Table 9 provides the results of these tests. Our results are generally in line with our main
-p
baseline results. Therefore, we believe that our findings are robust across different proxies for
re
cost of debt.
6.3. Board endogeneity treatments
lP
[Insert Table 9 here]
na
As with other studies on BOD characteristics (e.g., Elyasiani and Zhang, 2015; Chakravarty and Rutherford, 2017), our research is no exempt from the potential endogeneity
Jo ur
issues.9 We, therefore, perform several methods to control for this problem. We first employ the one-year lag of all independent variables (Table 10) and re-test them under OLS with robust standard errors.
We next perform Two-Stage Least Square (2SLS) (Table 10: Panel A) and Three-Stage Least Square (3SLS) (Table 11: Panel B) to reduce endogeneity between busy BOD and the bank cost of debt. Given that the focus of this study is on board busyness, we seek for good exogenous instrumental variable (IVs) for this main variable that is correlated with the suspected endogenous variable, but uncorrelated with the error terms of the dependent variable (Wooldridge, 2009). We followed the design of Elnahass et al. (2019) and Elyasiani and Zhang (2015) to choose IVs which include the number of public firms headquartered in
8
We aware of limitation of using this proxy for the cost of debt in IBs. However, in IBs, IAHs play an important role in providing capital to the banks; therefore, higher return for IAHs may imply higher cost of capital or equivalent to higher cost of debt. 9 Wu-Durbin-Hausmann statistics show an existence of endogeneity of board busyness measures.
18
Journal Pre-proof the same country of the bank and the country-level income-generating category10 (recorded in World Bank). For the first IV, we argue that directors of the banks headquartered in countries with more public companies are more likely to find additional jobs in other firms; hence, the number of outside directorships is predicted to be positively related to the number of public firms headquartered in the same country. For the second IV, we contend that a more developed economic system with high-income levels tends to feature skilled and high-paid jobs for directors (World Bank, 2016); thereby, we anticipate a higher number of outside directorships hold by directors in banks located in high-income countries. These two IVs are expected to affect board busyness measure, but are uncorrelated with the error term in the
of
main equations.11 More specifically, they are correlated with possible endogenous variables
ro
(BOD busyness), and should only predict bank cost of debt indirectly through their impacts on endogenous variables (see Black et al., 2006; Elnahass et al., 2019). For our estimated
-p
sample, we argue that these IVs employed in our study can indirectly influence bank cost of debt since the country-level indicators are less likely to affect individual banks’ cost of debt
re
endogenously. This is confirmed by two diagnostic tests, (i) The Sargan test (misspecification test with the null hypothesis of no misspecification) and (ii) the Breusch and Pagan LM test
lP
(to examine whether cross-equation disturbances are truly associated and if the equations need to be tested simultaneously). We treat either busy BOD or bank cost of debt as
na
endogenous variables and construct simultaneous equations models, Eqs (2) and (3). In the first equation, Eqs (2), we estimate the effect of busy directors on bank cost of debt; and in
Jo ur
the second equation, Eq. (3), we estimate the impact of such cost of debt on the busy BOD. Next, we use a two-step system Generalized Method of Moments (GMM) (Table 11, Panel C) estimator, which controls for the unobserved impacts through the transformation of the variables into first differences to reduce unobserved heterogeneity and omitted variable bias (Arellano and Bover 1995). GMM procedures employ lagged values as IVs for the endogenous variables such as board busyness (see Mollah and Zaman, 2015; Mollah et al., 2017). The logic is that these factors in earlier years could not have resulted from bank cost of debt in subsequent years; hence, endogeneity problem is unlikely.
10
It is measured as a dummy variable taking value of one if the “home” bank is in a country classified as a middle and highincome generating nation, and 0 otherwise. See more in Trinh et al. (2019). 11 With the purpose of checking the results sensitivities, we use alternative IV. Specifically, we take the year-average of the board busyness variable of other banks in the same country for our sample as IV. This method of instrumenting has been tested and verified by previous literature (e.g. John et al., 2008; Laeven and Levine 2009; Anginer et al., 2014; Safiullah and Shamsuddin, 2019). After using this instrument, the results suggest that the cost of debt variation is less likely to associate with board busyness in the other banks. Therefore, we can argue that this instrument is very likely to be correlated with the endogenous variable (board busyness) and at the same time be less likely to correlate with unobserved factors that affect dependent variables (i.e. cost of debts).
19
Journal Pre-proof
[Insert Table 10 and 11 here] Finally, we employ propensity score matching (PSM) estimation (Rosenbaum and Rubin, 1983; Elnahass et al., 2019) as a test of sample selection bias and possible endogeneity for board busyness variable. Three-step process is therefore as follows. Firstly, we employ probit technique to estimate propensity scores for banks having board with busy outside directors (treatment group) and those having board without busy outside directors (control group). Secondly, after obtaining estimated propensity scores of treated and controlled groups, we then continue matching samples using four alternative methods: one-to-one nearest neighbour
of
matching with and without replacement that the unit chosen from male directors
ro
observations, and nearest neighbour matching with n=2 and n=3 with replacement. Using these techniques can match each observation of treatment group with each observation of
-p
control group. The quality of matching is confirmed in appendix 1 demonstrating the
re
distribution of the propensity score before and after matching. Finally, we proceed to examine the average effects of board busyness on cost of debt through regressions on the
lP
matched samples only. Table 12 shows univariate (Panel A and B) and multivariate results (Panel C) for full sample. The former (i.e., average treatment effects in Panel A; average
na
treatment effects on the treated estimation with bootstrapping of standard errors (i.e. 100, 1000, 10000 replications) in Panel B) indicates that cost of debt is lower for treatment group,
Jo ur
or lower for banks having board with busy outside directors than their counterparts having board without busy outside directors. This is consistent across all employed matching techniques. The latter (i.e., regression results) shows negative and significant coefficients across all regressions 1-4. This suggests a negative relationship between board busyness and cost of debt for matched sample12, which is in line with our main results presented in Table 5.13 [Insert Table 12 here]
12
In unreported tests, we separate full sample into two subsamples of IBs and CBs, and then estimate PSM. We consistently find significantly negative effect of board busyness on cost of debt in CBs subsample, yet that finding is insignificant in IBs. This is in line with our results for each sub sample found in Table 6b. Tables are available upon request. 13 For our estimated period, there was no major regulatory requirements introduction related to board independence (Aljughaiman and Salama, 2019). We hence do not use difference-in-difference (DiD) type setting.
20
Journal Pre-proof Results obtained from all above endogenous treatment approaches (Lags form; GMM; 2SLS; and 3SLS; PSM) are reported in Table 10-12 (Panel A, B and C) confirm that our main results are robust and survive across different model specifications.14 7. Summary and conclusion The main purpose of this study is to examine the impact of board busyness on a bank’s cost of debt within mixed banking system (Islamic and Conventional). Extant literature shows competing findings on the role of busy directors play in corporate performance. Yet no studies have been conducted on investigating the link between busy directors and bank cost
of
of debt financing. Besides, using two global banking systems (Islamic and Conventional) as the setting, we examine the effect that busyness of board of directors has on a bank’s cost of
ro
debt.
-p
We first compliment the Chakravarty and Rutherford (2017) study by investigating the linkage between board busyness and the firm borrowing rates for the cross-country banking
re
sample. Based on this analysis, we find that board busyness has an inverse association with the bank cost of debt financing. Therefore, the higher the board busyness, the lower the firm
lP
cost of debt. While establishing the negative relationship between board busyness and the cost of debt for the whole sample, we also compare and contrast such relation between two
na
bank types with different governance mechanisms: one-layer (conventional banks) and twolayer (Islamic banks). We find that conventional banks having busy directors are more likely
Jo ur
to enjoy the low cost of debt than Islamic banks having busy directors. Furthermore, we find a non-linear effect of busy boards on borrowing rates in conventional banks, but not in their Islamic counterparts.
In this study, we extend the existing body of research on busy board of directors. Contrary to prevalent wisdom that board busyness tends to be detrimental to firm performance/value, we indicate that busy board can also provide a source of economic value to their respective banks and, hence, lowering the borrowing rate for their respective banks. By taking a modest step toward a more comprehensive understanding of busy outside directors for different banking models, we expect business practitioners, regulators and academics to gain a deeper understanding on how to balance the positive and negative attributes of board busyness as it associates with bank financing costs.
14
Unreported tests perform the propensity-score matching method. The results are consistent with the findings in main tests. Tables are provided upon request.
21
Journal Pre-proof *This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References
Jo ur
na
lP
re
-p
ro
of
Abdelsalam, O., Dimitropoulos, P., Elnahass, M., & Leventis, S. (2016). Earnings management behaviors under different monitoring mechanisms: The case of Islamic and conventional banks. Journal of Economic Behavior & Organization, 132, 155-173. Abedifar, P., Molyneux, P., & Tarazi, A. (2013). Risk in Islamic banking. Review of Finance, 17(6), 2035-2096. Aljughaiman, A., & Salama, A. (2019). Do banks effectively manage their risks? The role of risk governance in the MENA region. Journal of Accounting and Public Policy, 38(5), 106680. Alnasser, A. S., & Muhammed, J. (2012). Introduction to corporate governance from Islamic perspective. Humanomics, 28(3), 220-231. Anderson, R. C., Mansi, S. A., & Reeb, D. M. (2004). Board characteristics, accounting report integrity, and the cost of debt. Journal of Accounting and Economics, 37(3), 315-342. Anginer, D., Demirguc-Kunt, A., & Zhu, M. (2014). How does competition affect bank systemic risk?. Journal of Financial Intermediation, 23(1), 1-26. Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics, 68(1), 29-51. Bank for International Settlements (2017). Annual Report. Bebchuk, L. A., & Fried, J.M. (2004). Pay Without Performance: The Unfulfilled Promise of Executive Compensation. Harvard University Press, Cambridge, MA. Bebchuk, L. A., Fried, J. M., & Walker, D. I. (2002). Managerial power and rent extraction in the design of executive compensation (No. w9068). National bureau of economic research. Beck, T., Demirgüç-Kunt, A., & Merrouche, O. (2013). Islamic vs. Conventional Banking: Business Model, Efficiency and Stability. Journal of Banking & Finance, 37(2), 433–447. Black, B. S., Jang, H., & Kim, W. (2006). Does corporate governance predict firms' market values? Evidence from Korea. The Journal of Law, Economics, and Organization, 22(2), 366-413. Bliss, M. A., & Gul, F. A. (2012). Political connection and cost of debt: Some Malaysian evidence. Journal of Banking & Finance, 36(5), 1520-1527. Bloom, N., & Van Reenen, J. (2007). Measuring and explaining management practices across firms and countries. The Quarterly Journal of Economics, 122(4), 1351-1408. Boot, A. W., & Thakor, A. V. (1994). Moral hazard and secured lending in an infinitely repeated credit market game. International Economic Review, 899-920. Bradley, M., & Chen, D. (2011). Corporate governance and the cost of debt: Evidence from director limited liability and indemnification provisions. Journal of Corporate Finance, 17(1), 83-107. Cashman, G. D., Gillan, S. L., & Jun, C. (2012). Going overboard? On busy directors and firm value. Journal of Banking & Finance, 36(12), 3248-3259. Chakravarty, S., & Rutherford, L. G. (2017). Do busy directors influence the cost of debt? An examination through the lens of takeover vulnerability. Journal of Corporate Finance, 43, 429-443. Chen, J., & King, T. H. D. (2014). Corporate hedging and the cost of debt. Journal of Corporate Finance, 29, 221-245. Chen, Q., Filardo, A., He, D., & Zhu, F. (2016). Financial crisis, US unconventional monetary policy and international spillovers. Journal of International Money and Finance, 67, 62-81.
22
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Čihák, M., & Hesse, H. (2010). Islamic banks and financial stability: An empirical analysis. Journal of Financial Services Research, 38(2-3), 95-113. Cremers, K. M., Nair, V. B., & Wei, C. (2007). Governance mechanisms and bond prices. The Review of Financial Studies, 20(5), 1359-1388. Daher, M. (2017). Creditor control rights, capital structure, and legal enforcement. Journal of Corporate Finance, 44, 308-330. Deng, S. E., Elyasiani, E., & Mao, C. X. (2007). Diversification and the cost of debt of bank holding companies. Journal of Banking & Finance, 31(8), 2453-2473. Deng, S., Elyasiani, E., & Mao, C. X. (2017). Derivatives-hedging, risk allocation and the cost of debt: Evidence from bank holding companies. The Quarterly Review of Economics and Finance, 65, 114-127. Diamond, D. W. (1989). Reputation acquisition in debt markets. Journal of Political Economy, 97(4), 828-862. Easley, D., & O'hara, M. (2004). Information and the cost of capital. The Journal of Finance, 59(4), 1553-1583. Elnahass, M., Omoteso, K., Salama, A., & Trinh, V. Q. (2019). Differential market valuations of board busyness across alternative banking models. Review of Quantitative Finance and Accounting. Elyasiani, E., & Zhang, L. (2015). Bank holding company performance, risk, and “busy” board of directors. Journal of Banking & Finance, 60, 239-251. Ertugrul, M., & Hegde, S. (2008). Board compensation practices and agency costs of debt. Journal of Corporate Finance, 14(5), 512-531. Faleye, O., Hoitash, R., & Hoitash, U. (2011). The costs of intense board monitoring. Journal of Financial Economics, 101(1), 160-181. Fama, E. F. (1980). Agency problems and the theory of the firm. Journal of political economy, 88(2), 288-307. Fama, E. F., & Jensen, M. C. (1983). Separation of ownership and control. The Journal of Law and Economics, 26(2), 301-325. Ferris, S. P., Jagannathan, M., & Pritchard, A. C. (2003). Too busy to mind the business? Monitoring by directors with multiple board appointments. The Journal of finance, 58(3), 1087-1111. Fich, E. M., & Shivdasani, A. (2006). Are busy boards effective monitors? The Journal of finance, 61(2), 689-724. Fields, L. P., Fraser, D. R., & Subrahmanyam, A. (2012). Board quality and the cost of debt capital: The case of bank loans. Journal of Banking & Finance, 36(5), 1536-1547. Francis, J. R., Khurana, I. K., & Pereira, R. (2005). Disclosure incentives and effects on cost of capital around the world. The accounting review, 80(4), 1125-1162. Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305-360. John, K., Litov, L., & Yeung, B. (2008). Corporate governance and risk taking. The Journal of Finance, 63(4), 1679-1728. Kisgen, D. J., & Strahan, P. E. (2010). Do regulations based on credit ratings affect a firm's cost of capital? The Review of Financial Studies, 23(12), 4324-4347. Klock, M. S., Mansi, S. A., & Maxwell, W. F. (2005). Does corporate governance matter to bondholders? Journal of Financial and Quantitative Analysis, 40(4), 693-719. Laeven, L., & Levine, R. (2009). Bank governance, regulation and risk taking. Journal of Financial Economics, 93(2), 259-275. Lorca, C., Sánchez-Ballesta, J. P., & García-Meca, E. (2011). Board effectiveness and cost of debt. Journal of Business Ethics, 100(4), 613-631.
23
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Mollah, S., & Zaman, M. (2015). Shari’ah supervision, corporate governance and performance: Conventional vs. Islamic banks. Journal of Banking & Finance, 58, 418-435. Mollah, S., Hassan, M. K., Al Farooque, O., & Mobarek, A. (2017). The governance, risktaking, and performance of Islamic banks. Journal of Financial Services Research, 51(2), 195-219. Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial Economics, 5(2), 147-175. Pathan, S., & Faff, R. (2013). Does board structure in banks really affect their performance? Journal of Banking & Finance, 37(5), 1573-1589. Pathan, S., Wong, P. H., & Benson, K. (2019). How do ‘busy’and ‘overlap’directors relate to CEO pay structure and incentives?. Accounting & Finance, 59(2), 1341-1382. Penas, M. F., & Unal, H. (2004). Gains in bank mergers: Evidence from the bond markets. Journal of Financial Economics, 74(1), 149-179. Piot, C., & Missonier-Piera, F. (2009). Corporate governance reform and the cost of debt financing of listed French companies. Available at SSRN 960681. Rahaman, M. M., & Al Zaman, A. (2013). Management quality and the cost of debt: Does management matter to lenders? Journal of Banking & Finance, 37(3), 854-874. Reeb, D. M., Mansi, S. A., & Allee, J. M. (2001). Firm internationalization and the cost of debt financing: Evidence from non-provisional publicly traded debt. Journal of Financial and Quantitative Analysis, 36(3), 395-414. Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41-55. Safieddine, A. (2009). Islamic financial institutions and corporate governance: New Insights for Agency Theory. Corporate Governance: An International Review, 17: 142-158. Safiullah, M., & Shamsuddin, A. (2019). Risk-adjusted efficiency and corporate governance: Evidence from Islamic and conventional banks. Journal of Corporate Finance, 55, 105-140. Shivdasani, A. (1993). Board composition, ownership structure, and hostile takeovers. Journal of Accounting and Economics, 16(1-3), 167-198. Trinh, V. Q., Elnahass, M., Salama, A., & Izzeldin, M. (2019). Board busyness, performance and financial stability: does bank type matter? The European Journal of Finance, 1-28. Wintoki, M. B., Linck, J. S., & Netter, J. M. (2012). Endogeneity and the dynamics of internal corporate governance. Journal of Financial Economics, 105(3), 581-606. Wooldridge, J. M. (2009). On estimating firm-level production functions using proxy variables to control for unobservables. Economics Letters, 104(3), 112-114. World Bank (2016). “Labor market polarization in developing countries: challenges ahead.” Accessed 1 August 2017. http://blogs.worldbank.org. Yermack, D. (1996). Higher market valuation of companies with a small board of directors. Journal of Financial Economics, 40(2), 185-211.
24
Journal Pre-proof Table 1: Final Sample distributions for the Whole Sample Period Notes: The ultimate sample includes 70 listed banks (386 bank-year observations) with 27 IBs (150 bank-year observations) and 43 CBs (236 bank-year observations) for a six-year period from 2010. These banks are operating in 11 dual-banking countries. Country Observations Observations Observations % % % (IBs) (CBs) (Full Sample) (IBs) (CBs) (Full Sample) Bahrain 30 30 60 20.00 12.71 15.54 Bangladesh 36 44 80 24.00 18.64 20.73 Egypt 3 9 12 2.00 3.81 3.11 Indonesia 6 66 72 4.00 27.97 18.65 Jordan 12 29 41 8.00 12.29 10.62 Kuwait 3 12 15 2.00 5.09 3.89 Pakistan 24 6 30 16.00 2.54 7.77 Qatar 18 24 42 12.00 10.17 10.88 Saudi Arabia 6 6 12 4.00 2.54 3.11 UAE 6 6 12 4.00 2.54 3.11 Oman 6 4 10 4.00 1.70 2.59 Total bank-year observations 150 236 386 100 100 100 Number of banks 27 43 70 -
f o
l a
e
o r p
r P
n r u
o J
25
Journal Pre-proof Table 2: Variable Definitions and Sources Variables Abbreviations Cost of debt Cost of Debt
# Average directorships of directors % Busy directors Board of Directors Size Board Independence Chair-CEO duality Bank size Bank leverage GDP per capita Herfindahl-Hirschman Index Average country governance index Islamic banking dummy
Definitions After-tax weighted average cost of debt for the security. It is calculated by using government bond rates, a debt adjustment factor (DAF), the proportions of short-term and long-term debt to total debt, and the stock's effective tax rate. The debt adjustment factor represents the average yield above government bonds for a given rating class. Source: Bloomberg ABOD Average outside directorships per independent director, calculated as total number of additional (outside) boards held by independent directors divided by number of independent directors on the board. Source: annual report %BBOD Percentage of busy independent directors on the board (%), calculated as number of independent directors serving on two or more additional (outside) firms divided by number of independent directors on the board. Source: annual report LogBSIZE Natural logarithm of the total number of board of directors’ members. Source: annual report %IND Percentage of independent non-executive directors on the board of directors. Source: annual report DUAL Dummy variable which takes value of one if the chair and CEO is the same person and zero otherwise. Source: annual report LogTA Natural logarithm of the total assets. Source: DataStream LEV Total debt to total equity. Source: DataStream GDPCAPITA Natural logarithm of Gross domestic products per capita. Source: World Bank HHI The square of the sum of the ratio of total assets of each firm-year to total assets of all banks each year. Its range is from 0 to 1. Source: Bankscope GOV_COUNTRY The average value of six individual country governance measures including the corruption, government effectiveness, political stability, and regulatory quality, the rule of law, and voice and accountability. Source: World Bank ISLAMIC Dummy variable, taking value of one if the observed firm is classified as Islamic banks, otherwise zero. Source: Bankscope, annual report, and Central Bank’s website of each country.
f o
l a
e
o r p
r P
n r u
o J
26
Journal Pre-proof Table 3: Descriptive Statistics Notes: The table reports descriptive statistics of all variables employed in the main regression models of the study for the full sample. The ***, **, * represents p-values of 0.01, 0.05, and 0.10. See Table 2 and Section 6.1 for other variable definitions VARIABLES N Mean Median Std. Min Max Cost of Debt 386 2.281 1.394 2.532 0 10.320 ABOD 386 2.257 1.8 2.046 0 11 %BBOD 386 0.479 0.5 0.377 0 1 BSIZE 386 9.544 9 3.705 3 25 %IND 386 0.348 0.333333 0.237 0 1 DUAL 386 0.039 0 0.194 0 1 %INDQ 383 0.190 0 0.276 0 1 %INDEXP 386 0.294 0.2 0.339 0 1 LogASIZE 380 3.724 3 0.925 2 6 %BAC 380 0.495 0.5 0.333 0 1 INDMET 310 7.769 6 5.006 0 35 LogTA 386 15.408 15.427 1.287 11.999 18.047 LEV 386 8.039 7.775 3.774 -4.210 19.998 HHI 386 0.142 0.109 0.095 0.058 0.672 GDPCAPITA 386 8.750 8.216 1.542 6.634 11.480 GOV_COUNTRY 386 -0.286 -0.212 0.526 -1.181 0.737 ISLAMIC 386 0.389 0 0.488 0 1
f o
l a
e
o r p
r P
n r u
o J
27
Journal Pre-proof Table 4: Correlation Matrix Notes: The table presents the Pearson pair-wise correlation matrix significance at the 5% level. See Table 2 for variable definitions. (1) (2) (3) 1. ABOD 1 2. %BBOD 0.813* 1 3. LogBSIZE 0.128* 0.191* 1 4. %IND 0.165* 0.201* -0.403* 5. DUAL 0.078 0.064 0.060 6. LogTA -0.038 0.005 -0.039 7. LEV -0.080 -0.053 0.153* 8. GDPCAPITA 0.064 0.109* -0.153* 9. HHI 0.022 0.011 -0.024 10. GOV_COUNTRY 0.112* 0.127* -0.144* 11. ISLAMIC -0.072 -0.097 0.233*
among main independent variables used in our main analysis for full subsample. * indicates (4)
(5)
(6)
1 -0.035 0.218* -0.170* 0.260* -0.115* 0.323* -0.071
1 0.042 0.050 -0.070 0.019 -0.099 -0.160*
1 0.042 0.478* 0.071 0.430* -0.111*
l a
o J
28
f o
ro
-p
e r P
n r u
(7)
1 -0.449* -0.115* -0.489* 0.052
(8)
(9)
(10)
(11)
1 0.146* 0.630* -0.001
1 0.074 0.144*
1 -0.063
1
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 5: Cost of Debt and Board of Directors’ Busyness The table presents OLS regression results identifying the effect of busy boards of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt ABOD -0.353*** (0.000) ABOD*ISLAMIC 0.158** (0.045) %BBOD -1.623*** (0.000) %BBOD*ISLAMIC 0.763* (0.079) LogBSIZE -2.710*** -2.629*** (0.000) (0.000) %IND 1.685*** 1.802*** (0.001) (0.000) DUAL -1.985*** -2.111*** (0.000) (0.000) LogTA 0.587*** 0.598*** (0.000) (0.000) LEV -0.072** -0.071** (0.015) (0.015) GDPCAPITA -1.297*** -1.214*** (0.000) (0.000) HHI -3.471*** -3.601*** (0.000) (0.000) GOV_COUNTRY 1.910*** 1.673*** (0.000) (0.000) ISLAMIC -1.142*** -1.187*** (0.000) (0.000) Constant 13.000*** 11.826*** (0.000) (0.000) Year fixed effects YES YES Observations 386 386 R-squared 0.600 0.581
29
Journal Pre-proof Table 6a: Trend Analysis: Bank Cost of Debt by the Interval Categories of Board of Directors’ Busyness The table presents trend analysis results identifying a bank’s cost of debt for every incremental change in board busyness proxy measures. Panel A % busy outside directors Intervals <10% 10-19% 20-29% 30-39% 40-49% 50-59% 60-69% 70-79% 80-89% 90-99% N 117 1 15 17 8 62 41 38 7 0 % 30.3% 0.3% 3.9% 4.4% 2.1% 16.1% 10.6% 9.8% 1.8% 0 Cost of debt 2.842 0.659 0.313 0.096 0.089 0.010 1.916 2.790 3.267 Panel B Average number of outside directorships hold by outside directors Intervals <1 1-1.99 2-2.99 3-3.99 4-4.99 5-5.99 6-6.99 7-7.99 8-8.99 9-9.99 N 109 89 57 57 25 22 14 9 1 1 % 28.2% 23.1% 14.8% 14.8% 6.5% 5.7% 3.6% 2.3% 0.3% 0.3% Cost of debt 3.385 2.562 1.744 1.620 1.445 1.202 1.187 0.917 0.243 0.579
ro
f o
l a
p e
r P
n r u
o J
30
100% 80 20.7% 3.388 >9.99 2 0.4% 1.902
Journal Pre-proof Table 6b: Possible Non-linear Relationship between Cost of Debt and Board of Directors’ Busyness The table presents OLS regression results for the full sample (Panel A), conventional bank subsample (Panel B) and Islamic bank subsample (Panel C) identifying the possible nonlinear effect of busy board of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses.
VARIABLES ABOD (ABOD)2 %BBOD (%BBOD)2 LogBSIZE %IND DUAL LogTA LEV GDPCAPITA HHI GOV_COUNTRY ISLAMIC Constant Year fixed effects Observations
PANEL A: FULL SAMPLE (1) (2) -0.551*** (0.000) 0.035** (0.019) -2.402*** (0.010) 1.067 (0.217) -2.492*** -2.555*** (0.000) (0.000) 2.340*** 2.220*** (0.000) (0.000) -2.065*** -2.109*** (0.000) (0.000) 0.566*** 0.587*** (0.000) (0.000) -0.076*** -0.067** (0.010) (0.028) -1.192*** -1.183*** (0.000) (0.000) -3.662*** -3.782*** (0.000) (0.000) 1.625*** 1.577*** (0.000) (0.000) -0.840*** -0.840*** (0.000) (0.000) 11.800*** 11.361*** (0.000) (0.000) YES YES 386 386
PANEL C: ISLAMIC BANKS (5) (6) -0.028 (0.883) -0.006 (0.820) -0.031 (0.981) -0.326 (0.794) -2.338*** -2.297*** (0.000) (0.000) -0.341 -0.362 (0.522) (0.516) -
0.455*** (0.000) -0.011 (0.731) -0.869*** (0.003) -2.417** (0.012) 1.579** (0.036)
0.450*** (0.000) -0.008 (0.806) -0.840*** (0.002) -2.440*** (0.009) 1.493** (0.024)
12.437*** (0.000) YES 236
8.748*** (0.007) YES 150
8.450*** (0.006) YES 150
f o
l a
n r u
o J
PANEL B: CONVENTIONAL BANKS (3) (4) -0.768*** (0.000) 0.057*** (0.002) -4.312*** (0.000) 2.711** (0.019) -2.055*** -2.282*** (0.000) (0.000) 4.126*** 4.409*** (0.000) (0.000) -2.002*** -2.009*** (0.000) (0.000) 0.587*** 0.605*** (0.000) (0.000) -0.187*** -0.208*** (0.000) (0.000) -1.314*** -1.272*** (0.000) (0.000) -5.101*** -5.413*** (0.001) (0.000) 1.429*** 0.987* (0.005) (0.053)
o r p
e
r P
12.103*** (0.000) YES 236
31
Journal Pre-proof Adjusted R2
0.602
0.580
0.658
0.640
0.394
f o
l a
e
o r p
r P
n r u
o J
32
0.394
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 7: Cost of Debt and Board of Directors’ Busyness: The Effect of Agency Costs The table presents OLS regression results identifying the effect of agency costs on the relationship between busy boards of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt ABOD -0.350*** (0.000) ABOD*CASH -0.001 (0.996) ABOD*CASH* ISLAMIC 0.107*** (0.001) %BBOD -1.572*** (0.000) %BBOD*CASH 0.012 (0.961) %BBOD*CASH*ISLAMIC 0.765*** (0.000) LogBSIZE -2.607*** -2.556*** (0.000) (0.000) %IND 1.749*** 1.704*** (0.000) (0.000) DUAL -2.056*** -2.218*** (0.000) (0.000) LogTA 0.586*** 0.612*** (0.000) (0.000) LEV -0.047 -0.047* (0.101) (0.099) CASH -0.518*** -0.703*** (0.000) (0.000) GDPCAPITA -1.462*** -1.422*** (0.000) (0.000) HHI -2.664*** -2.521*** (0.000) (0.001) GOV_COUNTRY 2.559*** 2.462*** (0.000) (0.000) ISLAMIC -0.967*** -1.133*** (0.000) (0.000) Constant 14.551*** 13.814*** (0.000) (0.000) Year fixed effects YES YES Observations 386 386 R-squared 0.625 0.608
33
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 8: Sensitivity Test: Adding More Corporate Governance Variables The table presents OLS regression results identifying the effect of busy board of directors and other corporate governance variables on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 and Section 6.1 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt ABOD -0.374*** (0.000) ABOD*ISLAMIC 0.300*** (0.001) %BBOD -2.110*** (0.000) %BBOD*ISLAMIC 1.975*** (0.000) %INDQ 0.339 0.310 (0.360) (0.416) %INDEXP 0.298 0.182 (0.310) (0.557) LogASIZE 0.109 0.199 (0.440) (0.139) %BAC -0.850** -0.701** (0.011) (0.036) INDMET -0.012 -0.008 (0.601) (0.729) LogBSIZE -2.957*** -2.873*** (0.000) (0.000) %IND 0.540 0.618 (0.335) (0.291) DUAL -0.984** -1.226*** (0.036) (0.000) LogTA 0.698*** 0.666*** (0.000) (0.000) LEV -0.060* -0.061* (0.078) (0.071) GDPCAPITA -1.007*** -0.879*** (0.000) (0.000) HHI -3.907*** -3.868*** (0.000) (0.000) GOV_COUNTRY 1.321*** 1.084*** (0.002) (0.008) ISLAMIC -1.662*** -1.954*** (0.000) (0.000) Constant 9.663*** 8.566*** (0.000) (0.000) Year fixed effects YES YES Observations 310 310 R-squared 0.653 0.641
34
Journal Pre-proof Table 9: Alternative measures for Cost of Debt The table presents OLS regression results using alternative measures for cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust Pvalues in parentheses. See Table 2 for variable definitions. Panel A: Conventional Banks Panel B: Islamic Banks (1) (2) (3) (4) VARIABLES Interest Interest Return Return -0.355*** 0.768 ABOD (0.002)
(0.111)
5.067** (0.031)
0.803 (0.453) 3.452* (0.052)
Constant
YES 236 0.321
YES 236 0.344
YES 150 0.123
YES 150 0.156
Jo ur
na
lP
re
-p
ro
Year fixed effects Observations R-squared
of
1.432*** (0.000)
-0.541** (0.040) 0.843*** (0.000)
%BBOD
35
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 10: Endogeneity Treatment: Using one-year lag of independent variables The table presents OLS regression results identifying the effect of busy board of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt L.ABOD -0.391*** (0.000) L.ABOD*ISLAMIC 0.202** (0.027) L.%BBOD -1.979*** (0.000) L.%BBOD*ISLAMIC 1.212** (0.019) L.LogBSIZE -2.810*** -2.645*** (0.000) (0.000) L.%IND 1.673*** 1.809*** (0.001) (0.001) L.DUAL -2.118*** -2.249*** (0.000) (0.000) L.LogTA 0.616*** 0.624*** (0.000) (0.000) L.LEV -0.077** -0.081** (0.016) (0.011) L.GDPCAPITA -1.351*** -1.264*** (0.000) (0.000) L.HHI -3.134*** -2.989*** (0.000) (0.000) L.GOV_COUNTRY 1.992*** 1.780*** (0.000) (0.000) ISLAMIC -1.194*** -1.376*** (0.000) (0.000) Constant 12.555*** 11.293*** (0.000) (0.000) Year fixed effects YES YES Observations 316 316 R-squared 0.603 0.588
36
Journal Pre-proof Table 11: Robustness Check: Endogeneity Treatment by Two-stage Least Square, Three-stage Least Square and GMM The table presents 2SLS and 3SLS and GMM regression results identifying the effect of busy board of directors and other corporate governance variables on a bank’s cost of debt. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. Panel A: 2SLS Panel B: 3SLS Panel C: GMM VARIABLES (1) (2) (3) (4) (5) (6) Cost of Debt Cost of Debt Cost of Debt Cost of Debt Cost of Debt Cost of Debt ABOD -1.910*** -2.061*** -0.225*** (0.005) (0.000) (0.002) ABOD*ISLAMIC 1.711*** 1.865*** 0.195** (0.009) (0.000) (0.038) %BBOD -13.444*** -14.803*** -1.520*** (0.002) (0.000) (0.001) %BBOD*ISLAMIC 11.782*** 13.016*** 1.454** (0.003) (0.001) (0.041) LogBSIZE -1.090 0.551 -0.914 0.952 -0.946** -1.031 (0.271) (0.656) (0.178) (0.437) (0.017) (0.224) %IND 2.073** 4.055*** 2.095*** 4.319*** -0.497 0.533 (0.025) (0.004) (0.009) (0.001) (0.442) (0.593) DUAL -0.941 -1.197 -0.819 -1.059 -0.373 0.183 (0.476) (0.331) (0.336) (0.306) (0.336) (0.888) LogTA 0.345** 0.253 0.347** 0.248 0.275* 0.294 (0.021) (0.206) (0.033) (0.253) (0.065) (0.257) LEV -0.048 -0.082* -0.043 -0.080 -0.021 -0.027 (0.204) (0.096) (0.378) (0.190) (0.561) (0.658) GDPCAPITA -1.675*** -1.237*** -1.678*** -1.188*** -0.477*** -0.394* (0.000) (0.000) (0.000) (0.000) (0.004) (0.093) HHI 0.774 2.114 0.486 1.678 -1.582*** -1.196** (0.732) (0.415) (0.823) (0.566) (0.000) (0.017) GOV_COUNTRY 3.638*** 2.734*** 3.721*** 2.729*** 0.988** 0.700 (0.000) (0.004) (0.000) (0.006) (0.012) (0.213) ISLAMIC -5.119*** -7.458*** -5.495*** -8.140*** -0.568* -0.867** (0.002) (0.001) (0.000) (0.000) (0.074) (0.033) Constant 19.630*** 15.042*** 20.066*** 15.056*** 3.347* 3.719
f o
l a
e
o r p
r P
n r u
o J
37
Journal Pre-proof (0.000)
(0.000)
(0.000)
(0.000)
YES 2SLS 386 0.000 0.000 0.009
YES 2SLS 386 0.000 0.000 0.100
YES 3SLS 386 0.000
YES 3SLS 386 0.000
0.122 0.000
0.344 0.000
Cost of Debtt-1 Year fixed effects Model Observations Wald 2 (p-value) Endogeneity (p-value) Over identification (p-value) Hansen-Sargan test (p-value) Breusch-Pagan LM test (p-value) AR(1) p-value AR(2) p-value Hansen test (p-value)
l a
e
o r p
r P
n r u
o J
38
f o
(0.071) 0.786*** (0.000) YES GMM 316 0.000
(0.238) 0.766*** (0.000) YES GMM 316 0.000
0.000 0.440 0.202
0.000 0.320 0.328
Journal Pre-proof Table 12: Propensity score matching method: The effect of board busyness on the cost of debt (Full sample) Estimator: propensity-score matching Treatment model: probit Dependent variable: cost of debt Panel A: Average treatment effects with nearest neighbour Matching method Treated Control Δ S.E. 1:1 matching without replacement Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.954 3.211 -1.256*** 0.304 1:1 matching with replacement Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.679 3.245 -1.566*** 0.502 Nearest neighbour (n=2) Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.679 2.847 -1.168** 0.414 Nearest neighbour (n=3) Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.679 2.935 -1.257*** 0.399 Panel B: Average treatment effect on the treated with 1:1 nearest neighbour matching and bootstrapping of standard errors No of obs. Replications Observed (Δ) Bias S.E. 388 100 -1.548*** 0.185 0.369 388 1000 -1.548*** 0.209 0.342 388 10000 -1.548*** 0.211 0.342 Panel C: Regressions on matched samples (1) (2) (3) (4) Independent variables 1:1 matching 1:1 matching Nearest Nearest neighbour without replacement with replacement neighbour (n=3) (n=2) BOD Busyness Dummy -1.221*** -1.197*** -1.014*** -0.999*** (0.000) (0.000) (0.000) (0.000) Controls Yes Yes Yes Yes Constant 12.817*** 16.590*** 14.421*** 13.891***
f o
l a
e
o r p
r P
n r u
o J
39
T-stat -6.11 -4.14 -6.11 -3.12 -6.11 -2.82 -6.11 -3.15 T-stat -4.19 -4.53 -4.53
Journal Pre-proof (0.000) (0.000) (0.000) (0.000) Adjusted R-squared 0.588 0.579 0.520 0.541 Observations 306 464 344 356 Note: The table present the propensity score matching technique showing the results of the average treatment effects (ATE) and the average treatment effect on the treated (ATT) with 1:1 nearest neighbour matching and bootstrapping of standard errors. The ATE and ATT of board busyness on the cost of debt (Δ) is estimated as the difference between the mean changes of banks having board with busy outside directors (column “Treated”) and that of matched banks having board without busy outside directors (column “Non-treated”). P-value is presented in parentheses. Tstatistics based on standard errors are presented in final column. *** p<0.01, ** p<0.05, * p<0.1.
f o
l a
e
o r p
r P
n r u
o J
40
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Appendix 1: Unmatched versus Matched sample: propensity score
41
Journal Pre-proof
Highlights The study documents that busy boards of directors enhance
Jo ur
na
lP
re
-p
ro
of
a bank’s financing capacity by lowering its cost of debt. Such negative association is more pronounced in conventional banks than their Islamic counterparts. Results survive across alternative proxies for cost of debt and board busyness, and alternative model specifications and methodology. Our results contribute to understanding the indispensable role busy boards play in debt financing
42
Vu Quang Trinh, Abdullah Aljughaiman, Ngan Duong Cao PII:
S1057-5219(20)30116-2
DOI:
https://doi.org/10.1016/j.irfa.2020.101472
Reference:
FINANA 101472
To appear in:
International Review of Financial Analysis
Received date:
3 September 2019
Revised date:
24 February 2020
Accepted date:
26 February 2020
Please cite this article as: V.Q. Trinh, A. Aljughaiman and N.D. Cao, Fetching better deals from creditors: Board busyness, agency relationships and the bank cost of debt, International Review of Financial Analysis(2020), https://doi.org/10.1016/ j.irfa.2020.101472
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier.
Journal Pre-proof
Fetching Better Deals from Creditors: Board Busyness, Agency Relationships and the Bank Cost of Debt Vu Quang Trinh*, Newcastle University Business School, Newcastle University, United Kingdom
-p
ro
of
*Corresponding author. Email: [email protected] Postal address: Newcastle University Business School, 05 Barrack Road, Newcastle Upon Tyne, United Kingdom. Postcode: NE1 4SE
Abdullah Aljughaiman
re
School of Business, King Faisal University, Saudi Arabia
lP
Email: [email protected] Postal address: Al-Ahsaa 31982, KSA
na
Ngan Duong Cao
University of Bath School of Management, United Kingdom
Jo ur
Email: [email protected] Postal address: Claverton road, Bath, BA 2 7AY, UK.
Journal Pre-proof
ABSTRACT In a cross-country setting, we document that busy boards of directors (i.e., outside directors with multiple directorships) enhance a bank’s financing capacity by lowering its cost of debt, which is consistent with the signalling quality hypothesis. Our analysis further reveals that this negative association is more pronounced in conventional banks than their Islamic counterparts. Possibly owning to the distinctive governance structure and the complexity of the Islamic business model, which requires closer monitoring, Muslim debtholders might depreciate a busy board of directors as it is likely to associate with lower scrutinising effectiveness. Our results provide a positive counterpoint to the negative relationship that exists between busy directors and firm performance, and contributes to understanding the indispensable role busy boards play in debt financing.
-p
ro
of
Keywords: Board busyness, Islamic banks, Conventional banks, Cost of debt JEL classification: C23 G01 G21 G28 L50 M4
Jo ur
na
lP
re
Declarations of interest: none
Journal Pre-proof 1. Introduction “Given that bank loans are the primary source of financing for most economies, the cost of debt plays a significant role in an economy’s growth and performance” (Chen et al., 2016, p.70) Debt contracts are reflected as a central mechanism to economic growth because they provide a useful source to finance investments (Daher, 2017). Like other industrial firms, outside equity and debt remain the two primary finance sources of banks’ capital. Nevertheless, relatively to the former source, the debt market appears to be a predominant source of financing of banks, which is evident through its tremendous size1 and its large share
of
in a typical bank’s total assets. Hence, banks are well-acknowledged as relatively highly
ro
leveraged and heavily regulated as compared to non-financial firms. Given the importance of debt finance to the banking industry, it is worthy of special attention from academic
-p
researchers and practitioners.
Conventionally adhered to the agency theory, interest conflicts among the three parties
re
comprising of directors, shareholders, and debtholders are inevitable (Chakravarty and
lP
Rutherford, 2017). Particularly, shareholders of high-leveraged institutions, including banks often have a strong incentive to expropriate wealth from debtholders through investing in
na
risky projects (Fields et al., 2012). Debtholders may anticipate these incentives, and thus, require greater risk premium (returns) on their capital, leading to a higher cost of debt (see Jensen and Meckling, 1976; Ertugrul and Hegde, 2008). Creating a high-quality board of
Jo ur
directors (BOD hereafter) to scrutinise bank activities and investments on behalf of debtholders is a fundamental solution for such conflict of interest. Numerous existing studies (e.g., Anderson et al., 2004; Fields et al., 2012) have focused on the linkage between quantifiable measures of the management quality (i.e., corporate governance and board attributes) and various corporate outcomes such as cost of debt, equity return, and firm value. Some of them demonstrate the effect of the BOD characteristics on the cost of debt financing in non-financial firms (e.g., Anderson et al., 2004; Piot and MissonierPiera, 2007; Lorca et al., 2011). They generally claim that debtholders favour strong monitoring mechanisms, which tend to limit managerial opportunism. Thus, the quality of governance mechanisms can be considered as a channel, which may reduce creditors’ risk, and therefore, debtholders might become more lenient in the way the risk premium is set. In 1
Statistics of Bank for International Settlements (2017) show that at the end of the first quarter of 2017, the global debt securities outstanding reached to 21,749 billion of US dollars; among this, international debt securities outstanding of banks (6,427 billion of US dollars) are double than that of non-financial firms (3,261 billion of US dollars).
1
Journal Pre-proof other words, the increased monitoring effectiveness of the board towards the management team and CEO as a default risk-reduction corporate governance mechanism may benefit the debtholders as it aligns the interest conflicts between them and agents. This raises the high probability that the bank can meet the interest of the creditors and principal payment obligations. Although those studies help understand the impacts of management quality on different corporate outcomes, the research is still limited within the contexts of busy boards and cost of debts. As delegated monitors, outside directors are charged with legal roles and accountabilities offering effective oversight over executives or managers such that conflicts
of
of interest between creditors, shareholders and managers are minimised. Debtholders,
ro
therefore, can benefit from outside directors’ effective scrutinising through mitigation in agency problems such as shirking, perquisite consumption, and overinvestment (Trinh et al.,
-p
2019). Such decreased agency cost, together with the reduction in information asymmetry, can mitigate the cost of debt financing (Chen and King, 2014).
re
The existing literature studying cost of debt usually excludes banks from their samples, except for Penas and Unal (2004), Deng et al. (2007), Deng et al. (2017). However, those
lP
banking studies concentrate solely on the association between the borrowing rate and nongovernance aspects such as diversification (i.e., geographic, assets and non-traditional
na
activities) (e.g., Deng et al., 2007; Deng et al., 2017). Therefore, based on our review of literature, the association between the BOD busyness and cost of debt remains undiscovered,
Jo ur
specifically for financial firms.
As discussed earlier, debt financing is considered as the main source of funds in the banking industry. Thus, banks need to assure the achievement of an effective management to their debt terms. A highly qualified board in the banks can achieve this important goal. Fields et al. (2012) argue that the quality of the board might contribute to influence on the cost of debt capital. Taking in consideration the significant influence of busy board on the board quality, it is important to investigage the busy board in the context of banking industry. A board that is characterised as busy may affect the board quality positively through their experience and knowledge. This may lead to greater monitoring and better quality advice to the management, which may mitigate the banks cost of debt. Accordingly, our study extends the cross-country banking literature by answering the following important question: Does busy BOD reduce the cost of debt in the banking sector? The study brings critical values since debt financing is one of the crucial capital sources for bank operations (leveraged), and we are one of the first to look at this specific research 2
Journal Pre-proof question. While many prior studies show a negative effect of board busyness and firm performance, which is considered as problematic for the companies’ shareholders (e.g., Fich and Shivdasani, 2006; Cashman et al., 2012), there is very limited evidence showing that whether lenders prefer a firm’s board busyness or not (except for Chakravarty and Rutherford, 2017). Furthermore, since our sample covers countries that use dual banking operating systems (Conventional and Islamic), special attention has been paid to Islamic banks. Our focus on these two bank types is crucial under the on-going debate about the resilience and stability of the Islamic banks. Some researchers cast doubt on the long-run performance and risk-taking
of
of Islamic banks and underpin the need for further investigations for their governance
ro
structure and board efficacy (e.g., Cihák and Hesse, 2010; Beck et al., 2013; Abdelsalam et al., 2016). While conventional banks operate within the traditional interest-based framework,
-p
their Islamic counterparts refer to a banking system that operates under risk-sharing base. Specifically, Islamic banks must follow several Shari’ah principles that distinct their
re
operating system and have its own multi-layer governance system (Mollah and Zaman, 2015)2. BODs and managers in Islamic banks are required to adhere to Islamic principles of
lP
Shari’ah in fulfilling their missions to maximise shareholders’ wealth. The stakeholders’ interests in this bank type might extend beyond the financial interest to involve ethical and
na
religious values/needs (Alnasser and Muhammed, 2012). The failure to meet those expectations by the board and managers creates an extra source of agency problems. Given
Jo ur
these differences between conventional and Islamic banks, this paper is inspired by the demand to recognise the differences in busy board influences between two bank types. We argue that the need for monitoring of a BOD towards managers in Islamic banks tends to be higher than that in their conventional counterparts, leading to dissimilar effects of BOD busyness on cost of debt. We also believe that a comparative investigation of BOD busyness across two different bank types, Islamic and conventional banks, is particularly important to on-going debate on influences of board busyness on financial indicators when levels of agency and monitoring demands are different. This, in turn, can improve our understanding of busy directors’ role in unique bank governance. Therefore, we next aim to investigate whether the distinct features of Islamic banks alter the relationship between the cost of debt and busy board. 2
To be compliant with Sharia’ah Principles, Islamic banks are not allowed to charge interest on money, take excessive risks, invest in harmful activities (e.g., gambling, alcohol,), and their transaction must be based on real economic assets. Islamic banks financial instruments must be based on profit-losses sharing contracts.
3
Journal Pre-proof Our second research question is set upon on the argument that busy directors appear to have differential effects on bank performance depending on the level of agency conflicts related to the corporate operations (Trinh et al., 2019). Considering the monitoring and advising roles of directors, the value of holding multiple directorships depends ultimately on the relative importance of effective oversight, and the governance attributes. If the need for scrutinising is vital for certain types of firms, because of the greater business-process complexities and uniqueness of agency problems, the monitoring role of independent directors is relatively more important than their resource function. To investigate the distinct effects of board busyness on the cost of debt conditional on banks’ institutional
of
characteristics and business models, we evaluate such issue across two global banking
ro
systems: Islamic and conventional banks.
The data focuses on publicly traded commercial banks over the estimated period of 2010-
-p
2015. Several interesting results are obtained. First, we find a negative association between the banks’ cost of public debt and a busy BOD on our full sample. This confirms our main
re
hypothesis that busy BOD significantly reduces the cost of debt in banking sector. Second, the opposite signs of coefficients of interaction terms between busy boards and Islamic bank
lP
dummy factor suggest that the negative relationship between the cost of debt and a busy board is lessened for Islamic banking. In other words, by conditional on the bank types, we
na
find that Islamic banks having busy outside directors are less likely to enjoy lower cost of debt comparing to their conventional counterparts. However, we find a non-linear
Jo ur
relationship between busy BOD and cost of debt for conventional bank subsample, which is not significant in Islamic bank subsample. Our results are robust across different board busyness measures and various model specifications. Our findings contribute to the line of research that focuses on the relationship between board attributes and the cost of debt (e.g., Lorca et al., 2011; Chakravarty and Rutherford, 2017); and also complement comparative governance literature underpinning the Islamic conventional banking paradigm (e.g., Mollah and Zaman, 2015; Mollah et al., 20173; Trinh et al., 2019) by investigating the impact of the board busyness on a bank’s borrowing rates from an international perspective. For previous studies, while Mollah and Zaman (2015) examine a specific character of Shari’ah board (i.e. the board size) and its influence on bank performance, Mollah et al. (2017) continue investigating the relation between general
3
Mollah and Zaman (2015) examines a specific character of SSB (i.e. the board size of SSB) and its influence on bank performance. Mollah et al. (2017), investigates the relation between general corporate governance structures, via a developed index, and bank risk-taking across the two bank types.
4
Journal Pre-proof corporate governance structures, via a developed index, and bank risk-taking across the two bank types. More recently, Trinh et al. (2019) and Elnahass et al. (2019) have contributed to Mollah papers by testing the effect of board busyness (both BOD and Shari’ah board) on bank performance, risk-taking behaviour and market valuations. However, all these studies consider general corporate governance and/or busyness effects under the views of investors and banks. Our paper has therefore incrementally contributed to existing corporate governance research in Islamic and conventional banking system by comparatively evaluating board busyness function in a linkage with lenders or debtholders. This ends up with a consideration of bank cost of debt and board busyness within the specific context of
of
those both bank types.
ro
Furthermore, although Islamic banks are prohibited from paying an interest rate to their debtholders, they still pay an expected profit rate comparable to interest rates on other
-p
savings accounts. In this study, we also contribute to the method of measuring the cost of debt. While most of the existing studies on the cost of debt mostly focused on interest rates
re
when measuring a firm’s cost of debt (e.g. Reeb et al., 2001; Klock et al., 2005; Cremers et al., 2007; Kisgen and Strahan, 2010; Chen and King, 2014), this research uses a new
lP
alternative metrics: after-tax weighted average cost of debt (ATWACD). We argue that a bank often uses various sources, including either short- or long-term debt, to finance their
na
activities; therefore, using only interest rates as the cost of debt appears to fail in reflecting the lending price of a bank fully and accurately. ATWACD hence is more appropriate
Jo ur
because of its strong representation in either long-term or short-term debt. More importantly, our new measure is particularly useful in estimating the relative cost of debt of Islamic banks, which are known as interest-free businesses. Through this proxy for the cost of debt, analysis can reasonably evaluate whether and how debtholders value the busy boards of directors, especially those in the interest-free financial system like Islamic banks. The next section reviews the literature on the motivations for busy directors and the link between them and the cost of debt; and then sets the tested hypotheses. Section 3 describes our sample and database sources, and the methodology. Section 4 reports the descriptive statistics. Section 5 presents the empirical results, which is followed by additional testing and robustness checks (Section 6). Section 7 presents a conclusion. 2. Research background and empirical predictions 2.1. Sources of busy directors’ benefit: reduction in the cost of debt
5
Journal Pre-proof An effective bank BOD is one of the crucial components of strong firm governance, especially after recent financial turmoil. Nevertheless, the governing bodies of the BOD appear to provide limited regulation regarding specific guidance on how BOD members are anticipated to execute their role. Furthermore, the BOD functioning (e.g. BOD size, independence, directorships) is in general open to interpretation. For instance, the common question that if larger (or smaller), diverse (or uniform), and highly independent (or low independent) BOD within banks is more effective are still vague (e.g., Elyasiani and Zhang, 2015). Similarly, findings from present works are also equivocal with regards to the true nature of the actual role of busy directors (those are sitting in several for-profit firms’ boards)
of
on bank financial decisions (Elnahass et al., 2019). Advocates support that the rich sources
ro
brought by busy directors are beneficial to banks whilst opponents provide detrimental evidence.
-p
Corporate finance theorists assert that the value of a firm/bank is the present value of future cash flows (CFs) (Chen and King, 2014). Thus, the effects of busy BOD on bank value
re
can be from at least two sources: the impact on the CFs stream and/or the effect on the cost of capital that future CFs are discounted. As highlighted by Easley and O’Hara (2004), the
lP
fundamental in firm policies is the cost of capital due to its influence on profitability, and thereby investment decisions. Chakravarty and Rutherford (2017) is a notable recent work
na
providing empirical evidence for non-financial firms in a single country (US) through a hostile takeover framework that board busyness contributes to the reduced cost of debt.
Jo ur
However, the association of busy boards and the cost of debt, particularly in the banking industry, within previous research is still unanswered. Our study is different to Chakravarty and Rutherford by three-fold. First, we conduct a study on a global basis while their study is a single-country research. Second, we explore BOD busyness-cost of debt nexus in the context of banking. It is important to investigate such association in this sector as banks characteristics and business models are different from non-financial firms. Particularly, banks have high leverage level, encounter contagion risk in the industry, work in a significantly regulated environment, and have important impacts on the real economy (Elyasiani and Zhang, 2015). Furthermore, the banking industry is subject to unique agency costs that are different from the conventional agency problems (Safieddine, 2009). These different agency costs might distinguish the influence of busy board on the cost of debt in the banking industry. Consequently, we have addressed the lacking attention paid to the busy board of banks. Third, our study takes a step further to conduct a comparison research between two alternative banking models: conventional and Islamic banks. 6
Journal Pre-proof Existing corporate theories advise a close and important BOD busyness-borrowing costs nexus. In general, the theoretical foundation for the association between busy board and cost of debt could be driven from the managerial power theory (Bebchuk et al., 2002; Bebchuk and Fried, 2004). This is consistent with Pathan et al. (2019), who discuss that this theory recognises the important monitoring role that provided by the independent director. According to the managerial power theory, a firm with poor monitoring allow managers to acquire stronger power to benefit their interests that is relative related to the firm performance (Bebchuk et al., 2002; Bebchuk and Fried, 2004). This might give incentive to managers to take higher risk level, which maximise the shareholder wealth, on the expense of debt
of
holders. In addition, Fama (1980), Fama and Jensen (1983) argue that a busy board, that have
ro
a number of director memberships, can be an indication of his reputation as an effective monitoring of corporate managers. Taking together, the managerial power theory with this
-p
positive view, we can theoretically expect a negative relationship between busy board and cost of debt.
re
In term of empirical research, we thoroughly discuss below three main channels for how busy BODs could affect the bank cost of debt and develop the hypotheses accordingly.
lP
Our first arguable channel is related to weaker shareholder control. Weaker shareholders control could be created by having busy outside directors. The rationale is that directors held
na
additional directorship appointments tend to weaken shareholder control as they pursue conservation of their directorship reputation through a possible channel: putting anti-takeover
Jo ur
provisions in place (Chakravarty and Rutherford, 2017). A study by Cremers et al. (2007) suggests that not all board characteristics have the same influence on debtholders as they have on shareholders. For instance, shareholders prefer a strong shareholder control to maximise their wealth, while weak shareholder control is strongly desired by the debtholders to minimise the asset substitution problem. Thus, given such shareholder control-weakening behaviour of busy outside directors’ appeals to creditors, the benefits brought by those busy BODs are expected to outweigh the distraction risk to the bank stability (e.g. Faleye et al., 2011). This makes the lenders perceive busy boards positively, which may signal their preference by providing loans to firms at preferential rates (e.g. Shivdasani, 1993; Klock et al., 2005; Cremers et al., 2007). This supports a positive view that the number of other jobs holds by a “busy” directors may serve as a measure of his/her reputation as an effective monitor of corporate managers (Fama, 1980; Fama and Jensen, 1983), which diminishes managerial opportunism and mismanagement and hence, lower the cost of debt capital (Reeb et al., 2001; Ertugrul and Hegde, 2008). Furthermore, companies which protect their outside 7
Journal Pre-proof directors from a potential lawsuit, and hence allow these directors to pursue their self-interest at the expenses of shareholders, might enjoy a lower cost of debt (Bradley and Chen, 2011). Bradley and Chen (2011) also ascribe the association between the firm cost of debt and busy directors to the weaker shareholder control mechanism, because weaker shareholder control mitigates the asset substitution risk. Another argument that describes the association between the busy board and the cost of debt is through the insolvency risk channel. Trinh et al. (2019) find an inverse relationship between insolvency risks and busy outside directors within banks. Specifically, the probability of insolvency is considerably greater when BODs fail to fulfil their legal
of
responsibilities. Because multiple directorships help enhance outside directors’ ability,
ro
experience and networking to provide better advising and monitoring services for banks, insolvency risk is reduced. In particular, Elyasiani and Zhang (2015) find that busyness of
-p
directors improves their performance and consequently lowers the expected cost of financial distress. Wider networking of busy BOD directors helps banks to easier find good sources for
re
financing investments, leading to a reduction in the cost of borrowing. We hypothesise that busy BODs should be associated with a lower cost of debt. We further argue that the
lP
busyness of directors leads to a higher reduction on the cost of debt for banks with a higher insolvency risk. Banks with higher bankruptcy risk-benefit most from busy outside directors
na
due to their higher probability and costs of financial distress. In contrast, banks with little bankruptcy risk-benefit least from busy directors due to their unlikely default.
Jo ur
In a similar vein, companies with risky debt tend to forgo positive net present value investments if most value of the investment does toward the debtholders as poor states arise (Myers, 1977). Busy BODs can enhance the bank board advising and monitoring effectiveness through a decline of the probability of the poor states arising. Thereby, owners may have higher incentives as well as the confidence to invest their capital into potential value-enhancing projects (Elyasiani and Zhang, 2015). In other words, BOD busyness may lead the board following the optimal investment policy through making sufficient internal resources and enjoying a low cost of capital and hence, curtails the underinvestment problem when a bank has growth opportunities and a greater cost of external financing. In our particular case, banks having busy BODs may have a lower probability of the poor states occurring, and consequently, they should be the preference of the creditors. Based on the above, we reasonably contend that busy BODs should lead to a lower cost of debt by increasing the board effectiveness and mitigating the underinvestment problems.
8
Journal Pre-proof To sum up, our previous arguments indicate that effective BOD (e.g., busy board) can achieve high quality of management practices. Intuition contends that higher quality of management practices might fetch better deals from debtholders. Because BOD is a crucial input in the bank operating process, higher quality of BOD (better input) might lead to more efficient operations and thereby lower the bank underlying business risk. This, consequently, might relax the terms of credit the banks could get from their debtholders (Rahaman and Zaman, 2013). The simple economic framework adopted by Rahaman and Zaman (2013) can illustrate why the BOD management quality may substantially influence a bank’s borrowing cost via three points. First, although debtholders cannot observe the intrinsic quality of a
of
bank, the high-quality bank could always formulate costly signalling mechanisms/devices to
ro
fetch better deals from debtholders. The signal can be motivated in terms of reputational capital (Diamond, 1989) or relationship lending (Boot and Thakor, 1994) or by adopting
-p
better quality of management practices (Bloom and Van Reenen, 2007). All these motivations could be obtained from the multiple directorships of the directors. Second, debtholders may
re
not have sufficient motivation to fully scrutinise the bank-quality spectrum because large debtholders could always diversify away borrower-specific risks in their loan portfolios.
lP
Therefore, signalling by banks rather than monitoring by debtholders could be a better channel to understand the interaction between board quality and debt contract terms. Third,
na
the benefit from signals (adopting costly managerial practices) may not be sufficiently high for debtholders although those management quality practices enhance the firm total value
Jo ur
(Bloom and Van Reenen, 2007). It is therefore important to note that as long as signals from busy BODs are costly and benefits from signalling are sufficiently high, we should observe heterogeneity in BOD busyness across banks, and the variation in multiple directorships of BODs should also be correlated with the heterogeneity in the bank’s cost of debt. This argument is in line with Bloom and Van Reenen (2007) and Bloom et al. (2011) who provide agency explanations of why poorly-managed firms exist. Taken together, we hypothesise that board busyness is likely to reduce the bank cost of debt. Accordingly, we set our first hypothesis in the alternative forms: H1: Banks with busy boards exhibits lower cost of debt than banks with non-busy boards 2.2. Director busyness and the cost of debt: the matter of bank type Islamic banks differ primarily from conventional banks regarding the additional layer of governance of the former, which aims to overlook banks’ religious affairs. Particularly, the latter has no concerns on any religious matters whilst the former are required to conduct its 9
Journal Pre-proof activities in compliance with Islamic laws. In details, Islamic banks have to follow Shari’ah principles, which (1) prohibits payments or receive of interest on the debts (Riba), (2) relay on risk-sharing bases instead of risk-shifting for all financial transactions, (3) prohibits excessive risks and a certain type of investment activities (e.g., Alcoholic). To comply with these principles, the nature of Islamic banks’ liability side (e.g., financial contracts) has to be different from their counterparties (conventional banks). Even though interest in lending is prohibited in Islam, some of their transaction could be categorized as debt-based financing (Abedifar et al., 2013). In these financing contracts, the financer buys or ask to construct the underlying assets then sell it to the client using the deferred-payment basis with one or
of
several instalments. Another financing source that Islamic banks rely on is investment or
ro
saving accounts, which generates a return based on agreed profit rates. Abedifar et al. (2013) argue that the payoff to investment account holders in Islamic banks
-p
dependents on the bank’s performance and depositor’s religiosity. Religious depositors might accept lower return (cost of debt for Islamic banks) if the bank assures their compliance with
re
Shari’ah principles even if the bank's performance is poor. Therefore, their depositors might care more about the ethical and religious issues in the banks. In contrast, religious depositors
lP
might be risk-averse and demanding higher return if the bank showing a risky behaviour (Abedifar et al., 2013). Thus, the Islamic banks’ cost of debt can be affected negatively.
na
Given the unique characterises of Islamic banks and the additional restrictions in their business models, their BODs might encounter complicated and more responsibilities that
Jo ur
need to be achieved (e.g., compliant with Shari’ah principles). Aljughaiman and Salama (2019) discuss that the BOD responsibilities to comply with Shari’ah principles might lead to more constraints on their ability to monitor and advice management and manage risk, which might lead to differential impacts on their risk-taking behaviour. Thus, the differences in agency relations between Islamic and conventional banks can produce a different relationship between busy boards and cost of debt. Furthermore, board members are not Shari’ah scholars, and thus not able to confirm the bank's compliance with Shari’ah principles. Therefore, Islamic banks’ debtholders may not appreciate the busy board, and thus not have a strong influence on the cost of the fund they provide to the bank. In this study, we hypothesise that conventional banks employing busy BOD may enjoy lower cost of debt than Islamic banks with busy BOD. This is well explained by the finding of Trinh et al. (2019) which find that busy BODs in Islamic banks show detrimental effects on bank financial stability while those in conventional banks still bring reputational benefits
10
Journal Pre-proof such as networking and experience to the firms. This leads us to the second hypothesis as below: H2: Conventional Banks with busy boards exhibit lower cost of debt than Islamic banks with busy boards 3. Data and Research method 3.1. Data collection and sample To test our hypotheses, we employ several sources including the Thomson One Reuters DataStream, Bankscope, annual bank reports, and World Bank database.4 We focus on a
of
mixed sample of Islamic and conventional banks listed on stock markets from 2010 to 2015. This estimated period allows us to avoid the potential effect of the crisis shock of 2007-2009
ro
that banks encountered. We started with an initial sample of 3038 banks (196 Islamic banks and 2842 conventional banks) operating cross 36 countries. We then dropped any bank that
-p
have headquartered in the countries having only Islamic or conventional banks and having
re
less than two listed banks; dropped any other banks which do not have official websites and accounting period from 01 of January to 31 of December; excluded full investment banks and
lP
conventional banks with Islamic window (i.e., conventional banks with separate Islamic banking department with a supervision of Shari’ah board); and finally we do not include
na
banks which have less than three consecutive year’s full data availability. Our criteria are in line with the study of Elnahass et al. (2019) and other previous banking literature such as
Jo ur
Mollah et al. (2017).
Our ultimate dataset includes 386 firm-year observations (70 publicly quoted commercial banks operating in 11 countries). We report the sample distribution by country and bank in Table 1: 43 conventional banks (236 observations) and 27 Islamic banks (150 observations). We assume that the differences in Islamic banking models across countries due to various country-specific regulations do not exist. [Insert Table 1 here] 3.2. Research method
4
Consolidated bank accounting and market data were drawn from Datastream and Bankscope database. We manually collect corporate governance data (i.e. information of independent non-executive directors and the board of directors) from annual reports. We define a director as busy if he/she holds at least three directorships or two outside/additional directorships. We count the number of directorships held by directors in all for-profit private and public firms, consistent with previous studies (Elyasiani and Zhang, 2015; Chakravarty and Rutherford, 2017; Trinh et al., 2019). And finally, we collect macroeconomics indicators from the World Bank database.
11
Journal Pre-proof 3.2.1 Measuring agency cost of debt The primary dependent variable of our analysis is the bank cost of debt. We measure this by the after-tax weighted average cost of debt for the security, which is available in Bloomberg database. It is calculated by using government bond rates, a debt adjustment factor (DAF), the proportions of short-term and long-term debt to total debt, and the stock's effective tax rate. The debt adjustment factor represents the average yield above government bonds for a given rating class. The lower the rating, the higher the adjustment factor. The DAF is only employed when a bank does not have a fair market curve. When a bank does not have a credit rating, an assumed rate of 1.38 (the equivalent rate of a BBB+ Standard &
of
Poor's long-term currency issuer rating) is employed. The exact calculation of DAF is as
ro
follows:
-p
Cost of Debt = [[(SD/TD) * (CS * AF)] + [(LD/TD) * (CL * AF)]] * [1-TR] Where: SD is Short-term Debt. TD is total debt. CS is Pre-Tax Cost of Short-term Debt.
re
AF is Debt Adjustment Factor. LD is Long-term Debt. CL is Pre-Tax Cost of Long-term Debt.5 TR is Effective Tax Rate. Our measure is useful in estimating the relative cost of debt
lP
of Islamic banks, which are known as interest-free businesses. Accordingly, traditional interest rates are replaced by an equation, including government bond rates (short-term and
na
long-term debt) that both banking models can employ to finance for their operations.
Jo ur
3.2.2. Measuring the busyness of boards of directors A busy director is defined as an individual who serves in at least two outside firms. This definition is similarly employed in previous banking studies such as Elyasiani and Zhang (2015) and Trinh et al. (2019). Based on this, two measures of busy board of directors (BBOD) are used. The first measure is the ratio of outside directorships per outside director (ABOD), representing the average number of other outside board seats held by each outside director. It is computed as the total number of additional (outside) boards occupied by outside directors divided by the number of outside directors on the board (Ferris et al., 2003). The second measure is the percentage of busy outside directors (%BBOD), which is calculated as
5
In the case of Islamic bank, the cost of long-term debt is the return rate that the investment account holders receive from the bank.
12
Journal Pre-proof the number of outside directors serving on two or more additional (outside) firms divided by the number of outside directors on the board.6 3.2.3. Empirical models To examine our hypotheses, we employ the traditional ordinary least square (OLS) with robust standard error approach. Accordingly, we present the main equations model as follows: 𝐶𝑜𝑠𝑡 𝑜𝑓 𝐷𝑒𝑏𝑡𝑖𝑡 = 𝛽0 + 𝛽1 𝐵𝐵𝑂𝐷𝑖𝑡 + 𝛽2 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 + 𝛽3 𝐵𝐵𝑂𝐷𝑖𝑡 ∗ 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 + 𝜙𝑃 + µ𝑌𝑒𝑎𝑟 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝜀𝑖𝑡 𝐶𝑜𝑠𝑡 𝑜𝑓 𝐷𝑒𝑏𝑡𝑖𝑡 represents
the
bank
cost
of
Where,
(1) of
debt;
𝐵𝐵𝑂𝐷𝑖𝑡
represents
ro
{ABOD; %BBOD}; 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 represents the Islamic bank dummy variable, taking a value of
-p
1 if the observed bank is classified as Islamic, otherwise 0; 𝐵𝐵𝑂𝐷𝑖𝑡 ∗ 𝐼𝑆𝐿𝐴𝑀𝐼𝐶𝑖𝑡 is the interaction term between busy BOD and ISLAMIC dummy factor, which is used to control
re
for the mediating effects of the two different bank types; 𝜙𝑃 is a vector of control variables in the cost of debt model.
lP
Following prior studies (e.g., Lorca et al., 2011; Fields et al., 2012; Chakravarty and Rutherford, 2017), we control for board as well as firm-level and country-level characteristics
na
in our multivariate regressions to limit potential omitted variables bias. Those are widely employed in the previous cost of debt literature, which captures factors affecting the bank
Jo ur
borrowing rate. We first include a set of bank governance characteristics such as board size (LogBSIZE) and board independence (%INDEP) and CEO duality (DUAL). The former is measured by the number of directors on the board while the latter is measured by the percentage of outside non-executive directors on the board. The final is calculated as the dummy variable, which takes the value of one if the chair and CEO is the same person, otherwise, zero. We next control for bank and country-specific variables consisting of bank size (LogTA), bank leverage (LEV), GDP per capital (GDPCAPITA), Herfindahl-Hirschman index (HHI), and average index of country governance (GOV_COUNTRY). Table 2 contains a complete list of variable names and definitions/calculations. [Insert Table 2 here] 6
Due to limited data for Islamic banking, we are unable to distinguish between other types of directorships or activities of the director. We understand that this may be the limitation of the proxy of busyness, but this is consistent with previous studies in the field such as Elyasiani and Zhang (2015) and Trinh et al. (2019). In addition, in line with their studies, we have included only paid jobs, which affect significantly directors’ time and monitoring ability. In other words, we do not include voluntary jobs.
13
Journal Pre-proof 4. Descriptive statistics Table 3 provides descriptive statistics for the variables in all estimated models, which includes the cost of debt, busy board of directors, governance attributes, and other control firm-specific and country-specific variables. Our sample of 386 firm-years shows that the average and median cost of debt (Cost of Debt) are around 2.281% and 1.394%. Our t-test proves that the cost of debt of conventional banks - CBs (2.878%) is significantly higher than that of Islamic banks - IBs (1.343%). As for the board governance structure and board busyness, we find that, on average, the board of directors for the whole sample (IBs, CBs) composed of slightly fewer than 10
of
directors (11, 9) (BSIZE), which is consistent with the average number of directors on board
ro
in the study of Chakravarty and Rutherford (2017). An average director in our estimated sample holds about two outside directorship seats (ABOD), which is also similar to other
-p
previous research (Fich and Shivdasani, 2006; Cashman et al., 2012; Elyasiani and Zhang, 2015). This number is higher in CBs (2.374) compared to IBs (2.071); however, the t-test
re
indicates an insignificant difference between the two subsamples.
lP
Busy directors make up about 47.9% of a board of directors (%BBOD), 43.4% of an IB board and 50.85% of a CB board. We find that the percentage of busy directors on board in CBs is higher than that in IBs, showing that board of directors in listed CBs appears to be
na
busier than the board in listed IBs, supported by the result of two-sample t-test. In addition, we find that the percentage of independent directors (%INDEP) on the board for the full
Jo ur
sample (IBs, CBs) averagely is 35% (33%, 36%). [Insert Table 3 here]
Table 4 shows the results of the Pearson pair-wise correlation matrix for the main independent variables in our sample. We find that correlations among explanatory variables are within acceptable limits and raise no serious concerns on multicollinearity. [Insert Table 4 here] 5. Empirical findings 5.1. Busy board of directors and the cost of debt In Table 5 (Model 1 and 2), we show how the bank cost of debt is impacted when a bank adds an additional busy outside director to the board. We find that busy board of directors tends to reduce the bank cost of debt across two alternative proxy measures of board
14
Journal Pre-proof busyness (ABOD and %BBOD), after controlling for other board composition, firm-level and country-level characteristics. This finding is consistent with Chakravarty and Rutherford (2017), in their assertion that lenders recognise banks with busy independent directors as safer investments and are therefore willing to lend their capital at a lower interest rate. Additional studies show the similar concept include Shivdasani (1993), Klock et al. (2005), Cremers et al. (2007), Elyasiani and Zhang (2015), and so on. Economically, the coefficient of average directorships of outside directors suggests that for every additional directorship added to the board, the cost of debt is reduced by about 0.353%. In addition, the cost of debt for banks whose board increases 1% of busy outside
of
directors is reduced by about 1.623%. Based on these results, we conclude that board
ro
busyness has a negative and statistically significant impact on banks’ cost of debt, thus supporting our first hypothesis. Overall findings are consistent with our expectations that
-p
busy BODs tend to bring their reputational benefits to their banks (see Elnahass et al., 2019). This suggests that busy BODs are likely to make their banks more attractive under the eyes of
re
the debtholders, resulting in lower borrowing costs.
We also use interaction terms between busy BODs and IB dummy variable (i.e.
lP
ABOD*ISLAMIC; %BBOD*ISLAMIC) to conduct comparative analyses between IBs and CBs for the impact of busy outside directors on bank cost of debt; our second hypothesis.
na
Results generally show that busy BOD, across two alternative metrics (ABOD, %BBOD), has a significantly higher agency cost of debt in IB system than their conventional counterparts.
Jo ur
This is presented by the highly significant and positive coefficients of interaction terms in both models. This is in line with our expectations and our second hypothesis that CBs having busy BOD is likely to have a lower cost of debt than IBs having busy BOD.7 According to the magnitudes of the estimates of both interaction terms, it is possible for Islamic banks to still benefit from the busy boards for lower cost of debts. However, the reduced levels are lowered by approximately 50% (that results in coefficient equal to -0.195) comparing to those of conventional banks (-0.353) for ABOD. The finding implies that bank debtholders tend to be more concern about the busyness effect (e.g. lax attention, time and efforts) in IBs than that in CBs due to the higher complication of IBs in terms of agency relationships and institutional environments. [Insert Table 5 here] 7
In unreported tests, we run tests for the IB subsample separately by including size of Shari’ah board and multiple directorships of Shari’ah board. The findings are still consistent to our main results.
15
Journal Pre-proof The signs of other control variables are consistent with previous literature. For example, board size (LogBSIZE) is significantly and negatively related to the cost of debt, which suggests that banks having large BOD are likely to reduce the firm cost of debt. This is in line with previous studies of Lorca et al. (2011), Fields et al. (2012) and Chakravarty and Rutherford (2017) showing that larger board is negatively associated with firm borrowing rates. Results of board independence indicate that a high presentation of outside directors in boards is likely to increase the bank cost of debt significantly. This can be explained that banks with higher board independence tend to reduce the firm performance (Yermark, 1996, Wintoki et al., 2012; Pathan and Faff, 2013) and thus, the debtholders may require higher
of
premium from their lending.
ro
5.2. Trend analysis: bank cost of debt by the interval categories of board busyness
-p
We next follow the study of Chakravarty and Rutherford (2017) to provide a trend analysis of a bank average cost of debt for every incremental change in BOD busyness
re
variable. We indeed partition our board busyness proxy into equal increments and investigating how the cost of debt change with each increasing increment category.
lP
Table 6a (Panel A) reveal that we have split %BBOD (i.e. the percentage of busy outside directors serving on the BOD) into 10% segment categories: below 10%, 10-19.99%, 20-
na
29.99%, etc. Results show that at lower board busyness interval (up to 59.99%), the average cost of debt reduces as the percentage of busy outside directors increases. However, an
Jo ur
opposite result is found when the busyness interval exceeds 59.99%. Similarly, Table 6a (Panel B) provide findings on how the bank cost of debt is influenced when a bank adds an additional average outside directorships to the BOD. We explore that with every incremental addition of an average outside directorship to the BOD, the bank cost of debt decreases. However, the cost of debt slightly increases when the average outside directorships exceeds 8.99. These findings are in accord with prior studies in non-financial firms (e.g., Chakravarty and Rutherford, 2017) and in line with our main results. [Insert Table 6a here] 5.3. The possible non-linear effect of busy boards of directors on bank cost of debt We continue testing the possible non-linear effect of busy BOD on bank cost of debt between three subsamples: full sample, CBs and IBs subsample. We find for full sample and
16
Journal Pre-proof CBs subsample that, in Table 6b (Panel A and B, respectively), busy BOD is significantly related to a reduction in bank cost of debt. However, the sign directions of the square of board busyness, i.e. (ABOD)2 and (%BBOD)2 are reversed, implying a non-linear association between busy BOD and the bank cost of debt. This is consistent with the main result in Table 5 and the trend analysis in Section 5.2. We find no significant evidence for the effect of busy BOD on cost of debt in IBs. This finding, again, confirms our second hypothesis that CBs with busy BODs have a lower cost of debt than IBs with busy BODs. [Insert Table 6b here]
of
5.4. Agency relationships, busy board of directors and the cost of debt
ro
Trinh et al. (2019) argue that the agency problem of a bank could be affected by busy BOD through its functions of scrutinising, either positive or negative. We, hence, examine
-p
whether agency costs have any impacts on the main result related to board busyness-cost of debt nexus.
re
In Models 1 and 2 of Table 7, we report the results about the effect of agency costs (i.e.
lP
cash to total assets) on the relationship between busy BOD and bank cost of debt. We find that holding more cash in CBs appears not to show any significant influence on the negative
na
nexus between busy BOD and the borrowing rates. Meanwhile, holding higher cash level in IBs tends to increase such type of costs, which can be explained that debtholders in this bank
Jo ur
type seem not to have trust in the ability of busy outside directors in monitoring the cash use of managers. The findings are consistent with the study of Elnahass et al. (2019) showing that busy BODs in IBs are related to higher agency conflicts. [Insert Table 7 here]
6. Additional Testing and robustness checks 6.1. Effects of busy boards on cost of debt after controlling for additional control variables In this section, we test whether our main results in Table 5 are sensitive to changes in the components of control variables. We add additional independent variables related to corporate governance characteristics comprising of the qualifications of outside directors (i.e. %INDQ), the percentage of outside directors with financial expertise (i.e. %INDEXP). The size of audit committee in natural logarithm form (i.e. LogASIZE) the percentage of busy directors serving in audit committee (i.e. %BAC), and the average number of board meeting
17
Journal Pre-proof held by outside directors (i.e. INDMET). We find in Table 8 that the main findings remain unchanged, implying that our stories are robust when adding more control variables. [Insert Table 8 here] 6.2. Alternative measures of Cost of Debts For more robustness check, we use alternative measures for cost of debts. In details, we use the interest rate on the banks’ debt (i.e., Interest), which is computed as the interest expense for the year divided by the interest-bearing debt (Francis et al., 2005; Bliss and Gul, 2012), as a measure of cost of debt for CBs subsample. However, for the IB subsample, we
of
used the return on IAHs (i.e., Return)8 as an alternative proxy for its cost of debt because IBs
ro
are interest-free businesses. We derived these data from DataStream, Osiris and Bloomberg. Table 9 provides the results of these tests. Our results are generally in line with our main
-p
baseline results. Therefore, we believe that our findings are robust across different proxies for
re
cost of debt.
6.3. Board endogeneity treatments
lP
[Insert Table 9 here]
na
As with other studies on BOD characteristics (e.g., Elyasiani and Zhang, 2015; Chakravarty and Rutherford, 2017), our research is no exempt from the potential endogeneity
Jo ur
issues.9 We, therefore, perform several methods to control for this problem. We first employ the one-year lag of all independent variables (Table 10) and re-test them under OLS with robust standard errors.
We next perform Two-Stage Least Square (2SLS) (Table 10: Panel A) and Three-Stage Least Square (3SLS) (Table 11: Panel B) to reduce endogeneity between busy BOD and the bank cost of debt. Given that the focus of this study is on board busyness, we seek for good exogenous instrumental variable (IVs) for this main variable that is correlated with the suspected endogenous variable, but uncorrelated with the error terms of the dependent variable (Wooldridge, 2009). We followed the design of Elnahass et al. (2019) and Elyasiani and Zhang (2015) to choose IVs which include the number of public firms headquartered in
8
We aware of limitation of using this proxy for the cost of debt in IBs. However, in IBs, IAHs play an important role in providing capital to the banks; therefore, higher return for IAHs may imply higher cost of capital or equivalent to higher cost of debt. 9 Wu-Durbin-Hausmann statistics show an existence of endogeneity of board busyness measures.
18
Journal Pre-proof the same country of the bank and the country-level income-generating category10 (recorded in World Bank). For the first IV, we argue that directors of the banks headquartered in countries with more public companies are more likely to find additional jobs in other firms; hence, the number of outside directorships is predicted to be positively related to the number of public firms headquartered in the same country. For the second IV, we contend that a more developed economic system with high-income levels tends to feature skilled and high-paid jobs for directors (World Bank, 2016); thereby, we anticipate a higher number of outside directorships hold by directors in banks located in high-income countries. These two IVs are expected to affect board busyness measure, but are uncorrelated with the error term in the
of
main equations.11 More specifically, they are correlated with possible endogenous variables
ro
(BOD busyness), and should only predict bank cost of debt indirectly through their impacts on endogenous variables (see Black et al., 2006; Elnahass et al., 2019). For our estimated
-p
sample, we argue that these IVs employed in our study can indirectly influence bank cost of debt since the country-level indicators are less likely to affect individual banks’ cost of debt
re
endogenously. This is confirmed by two diagnostic tests, (i) The Sargan test (misspecification test with the null hypothesis of no misspecification) and (ii) the Breusch and Pagan LM test
lP
(to examine whether cross-equation disturbances are truly associated and if the equations need to be tested simultaneously). We treat either busy BOD or bank cost of debt as
na
endogenous variables and construct simultaneous equations models, Eqs (2) and (3). In the first equation, Eqs (2), we estimate the effect of busy directors on bank cost of debt; and in
Jo ur
the second equation, Eq. (3), we estimate the impact of such cost of debt on the busy BOD. Next, we use a two-step system Generalized Method of Moments (GMM) (Table 11, Panel C) estimator, which controls for the unobserved impacts through the transformation of the variables into first differences to reduce unobserved heterogeneity and omitted variable bias (Arellano and Bover 1995). GMM procedures employ lagged values as IVs for the endogenous variables such as board busyness (see Mollah and Zaman, 2015; Mollah et al., 2017). The logic is that these factors in earlier years could not have resulted from bank cost of debt in subsequent years; hence, endogeneity problem is unlikely.
10
It is measured as a dummy variable taking value of one if the “home” bank is in a country classified as a middle and highincome generating nation, and 0 otherwise. See more in Trinh et al. (2019). 11 With the purpose of checking the results sensitivities, we use alternative IV. Specifically, we take the year-average of the board busyness variable of other banks in the same country for our sample as IV. This method of instrumenting has been tested and verified by previous literature (e.g. John et al., 2008; Laeven and Levine 2009; Anginer et al., 2014; Safiullah and Shamsuddin, 2019). After using this instrument, the results suggest that the cost of debt variation is less likely to associate with board busyness in the other banks. Therefore, we can argue that this instrument is very likely to be correlated with the endogenous variable (board busyness) and at the same time be less likely to correlate with unobserved factors that affect dependent variables (i.e. cost of debts).
19
Journal Pre-proof
[Insert Table 10 and 11 here] Finally, we employ propensity score matching (PSM) estimation (Rosenbaum and Rubin, 1983; Elnahass et al., 2019) as a test of sample selection bias and possible endogeneity for board busyness variable. Three-step process is therefore as follows. Firstly, we employ probit technique to estimate propensity scores for banks having board with busy outside directors (treatment group) and those having board without busy outside directors (control group). Secondly, after obtaining estimated propensity scores of treated and controlled groups, we then continue matching samples using four alternative methods: one-to-one nearest neighbour
of
matching with and without replacement that the unit chosen from male directors
ro
observations, and nearest neighbour matching with n=2 and n=3 with replacement. Using these techniques can match each observation of treatment group with each observation of
-p
control group. The quality of matching is confirmed in appendix 1 demonstrating the
re
distribution of the propensity score before and after matching. Finally, we proceed to examine the average effects of board busyness on cost of debt through regressions on the
lP
matched samples only. Table 12 shows univariate (Panel A and B) and multivariate results (Panel C) for full sample. The former (i.e., average treatment effects in Panel A; average
na
treatment effects on the treated estimation with bootstrapping of standard errors (i.e. 100, 1000, 10000 replications) in Panel B) indicates that cost of debt is lower for treatment group,
Jo ur
or lower for banks having board with busy outside directors than their counterparts having board without busy outside directors. This is consistent across all employed matching techniques. The latter (i.e., regression results) shows negative and significant coefficients across all regressions 1-4. This suggests a negative relationship between board busyness and cost of debt for matched sample12, which is in line with our main results presented in Table 5.13 [Insert Table 12 here]
12
In unreported tests, we separate full sample into two subsamples of IBs and CBs, and then estimate PSM. We consistently find significantly negative effect of board busyness on cost of debt in CBs subsample, yet that finding is insignificant in IBs. This is in line with our results for each sub sample found in Table 6b. Tables are available upon request. 13 For our estimated period, there was no major regulatory requirements introduction related to board independence (Aljughaiman and Salama, 2019). We hence do not use difference-in-difference (DiD) type setting.
20
Journal Pre-proof Results obtained from all above endogenous treatment approaches (Lags form; GMM; 2SLS; and 3SLS; PSM) are reported in Table 10-12 (Panel A, B and C) confirm that our main results are robust and survive across different model specifications.14 7. Summary and conclusion The main purpose of this study is to examine the impact of board busyness on a bank’s cost of debt within mixed banking system (Islamic and Conventional). Extant literature shows competing findings on the role of busy directors play in corporate performance. Yet no studies have been conducted on investigating the link between busy directors and bank cost
of
of debt financing. Besides, using two global banking systems (Islamic and Conventional) as the setting, we examine the effect that busyness of board of directors has on a bank’s cost of
ro
debt.
-p
We first compliment the Chakravarty and Rutherford (2017) study by investigating the linkage between board busyness and the firm borrowing rates for the cross-country banking
re
sample. Based on this analysis, we find that board busyness has an inverse association with the bank cost of debt financing. Therefore, the higher the board busyness, the lower the firm
lP
cost of debt. While establishing the negative relationship between board busyness and the cost of debt for the whole sample, we also compare and contrast such relation between two
na
bank types with different governance mechanisms: one-layer (conventional banks) and twolayer (Islamic banks). We find that conventional banks having busy directors are more likely
Jo ur
to enjoy the low cost of debt than Islamic banks having busy directors. Furthermore, we find a non-linear effect of busy boards on borrowing rates in conventional banks, but not in their Islamic counterparts.
In this study, we extend the existing body of research on busy board of directors. Contrary to prevalent wisdom that board busyness tends to be detrimental to firm performance/value, we indicate that busy board can also provide a source of economic value to their respective banks and, hence, lowering the borrowing rate for their respective banks. By taking a modest step toward a more comprehensive understanding of busy outside directors for different banking models, we expect business practitioners, regulators and academics to gain a deeper understanding on how to balance the positive and negative attributes of board busyness as it associates with bank financing costs.
14
Unreported tests perform the propensity-score matching method. The results are consistent with the findings in main tests. Tables are provided upon request.
21
Journal Pre-proof *This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References
Jo ur
na
lP
re
-p
ro
of
Abdelsalam, O., Dimitropoulos, P., Elnahass, M., & Leventis, S. (2016). Earnings management behaviors under different monitoring mechanisms: The case of Islamic and conventional banks. Journal of Economic Behavior & Organization, 132, 155-173. Abedifar, P., Molyneux, P., & Tarazi, A. (2013). Risk in Islamic banking. Review of Finance, 17(6), 2035-2096. Aljughaiman, A., & Salama, A. (2019). Do banks effectively manage their risks? The role of risk governance in the MENA region. Journal of Accounting and Public Policy, 38(5), 106680. Alnasser, A. S., & Muhammed, J. (2012). Introduction to corporate governance from Islamic perspective. Humanomics, 28(3), 220-231. Anderson, R. C., Mansi, S. A., & Reeb, D. M. (2004). Board characteristics, accounting report integrity, and the cost of debt. Journal of Accounting and Economics, 37(3), 315-342. Anginer, D., Demirguc-Kunt, A., & Zhu, M. (2014). How does competition affect bank systemic risk?. Journal of Financial Intermediation, 23(1), 1-26. Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics, 68(1), 29-51. Bank for International Settlements (2017). Annual Report. Bebchuk, L. A., & Fried, J.M. (2004). Pay Without Performance: The Unfulfilled Promise of Executive Compensation. Harvard University Press, Cambridge, MA. Bebchuk, L. A., Fried, J. M., & Walker, D. I. (2002). Managerial power and rent extraction in the design of executive compensation (No. w9068). National bureau of economic research. Beck, T., Demirgüç-Kunt, A., & Merrouche, O. (2013). Islamic vs. Conventional Banking: Business Model, Efficiency and Stability. Journal of Banking & Finance, 37(2), 433–447. Black, B. S., Jang, H., & Kim, W. (2006). Does corporate governance predict firms' market values? Evidence from Korea. The Journal of Law, Economics, and Organization, 22(2), 366-413. Bliss, M. A., & Gul, F. A. (2012). Political connection and cost of debt: Some Malaysian evidence. Journal of Banking & Finance, 36(5), 1520-1527. Bloom, N., & Van Reenen, J. (2007). Measuring and explaining management practices across firms and countries. The Quarterly Journal of Economics, 122(4), 1351-1408. Boot, A. W., & Thakor, A. V. (1994). Moral hazard and secured lending in an infinitely repeated credit market game. International Economic Review, 899-920. Bradley, M., & Chen, D. (2011). Corporate governance and the cost of debt: Evidence from director limited liability and indemnification provisions. Journal of Corporate Finance, 17(1), 83-107. Cashman, G. D., Gillan, S. L., & Jun, C. (2012). Going overboard? On busy directors and firm value. Journal of Banking & Finance, 36(12), 3248-3259. Chakravarty, S., & Rutherford, L. G. (2017). Do busy directors influence the cost of debt? An examination through the lens of takeover vulnerability. Journal of Corporate Finance, 43, 429-443. Chen, J., & King, T. H. D. (2014). Corporate hedging and the cost of debt. Journal of Corporate Finance, 29, 221-245. Chen, Q., Filardo, A., He, D., & Zhu, F. (2016). Financial crisis, US unconventional monetary policy and international spillovers. Journal of International Money and Finance, 67, 62-81.
22
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Čihák, M., & Hesse, H. (2010). Islamic banks and financial stability: An empirical analysis. Journal of Financial Services Research, 38(2-3), 95-113. Cremers, K. M., Nair, V. B., & Wei, C. (2007). Governance mechanisms and bond prices. The Review of Financial Studies, 20(5), 1359-1388. Daher, M. (2017). Creditor control rights, capital structure, and legal enforcement. Journal of Corporate Finance, 44, 308-330. Deng, S. E., Elyasiani, E., & Mao, C. X. (2007). Diversification and the cost of debt of bank holding companies. Journal of Banking & Finance, 31(8), 2453-2473. Deng, S., Elyasiani, E., & Mao, C. X. (2017). Derivatives-hedging, risk allocation and the cost of debt: Evidence from bank holding companies. The Quarterly Review of Economics and Finance, 65, 114-127. Diamond, D. W. (1989). Reputation acquisition in debt markets. Journal of Political Economy, 97(4), 828-862. Easley, D., & O'hara, M. (2004). Information and the cost of capital. The Journal of Finance, 59(4), 1553-1583. Elnahass, M., Omoteso, K., Salama, A., & Trinh, V. Q. (2019). Differential market valuations of board busyness across alternative banking models. Review of Quantitative Finance and Accounting. Elyasiani, E., & Zhang, L. (2015). Bank holding company performance, risk, and “busy” board of directors. Journal of Banking & Finance, 60, 239-251. Ertugrul, M., & Hegde, S. (2008). Board compensation practices and agency costs of debt. Journal of Corporate Finance, 14(5), 512-531. Faleye, O., Hoitash, R., & Hoitash, U. (2011). The costs of intense board monitoring. Journal of Financial Economics, 101(1), 160-181. Fama, E. F. (1980). Agency problems and the theory of the firm. Journal of political economy, 88(2), 288-307. Fama, E. F., & Jensen, M. C. (1983). Separation of ownership and control. The Journal of Law and Economics, 26(2), 301-325. Ferris, S. P., Jagannathan, M., & Pritchard, A. C. (2003). Too busy to mind the business? Monitoring by directors with multiple board appointments. The Journal of finance, 58(3), 1087-1111. Fich, E. M., & Shivdasani, A. (2006). Are busy boards effective monitors? The Journal of finance, 61(2), 689-724. Fields, L. P., Fraser, D. R., & Subrahmanyam, A. (2012). Board quality and the cost of debt capital: The case of bank loans. Journal of Banking & Finance, 36(5), 1536-1547. Francis, J. R., Khurana, I. K., & Pereira, R. (2005). Disclosure incentives and effects on cost of capital around the world. The accounting review, 80(4), 1125-1162. Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305-360. John, K., Litov, L., & Yeung, B. (2008). Corporate governance and risk taking. The Journal of Finance, 63(4), 1679-1728. Kisgen, D. J., & Strahan, P. E. (2010). Do regulations based on credit ratings affect a firm's cost of capital? The Review of Financial Studies, 23(12), 4324-4347. Klock, M. S., Mansi, S. A., & Maxwell, W. F. (2005). Does corporate governance matter to bondholders? Journal of Financial and Quantitative Analysis, 40(4), 693-719. Laeven, L., & Levine, R. (2009). Bank governance, regulation and risk taking. Journal of Financial Economics, 93(2), 259-275. Lorca, C., Sánchez-Ballesta, J. P., & García-Meca, E. (2011). Board effectiveness and cost of debt. Journal of Business Ethics, 100(4), 613-631.
23
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Mollah, S., & Zaman, M. (2015). Shari’ah supervision, corporate governance and performance: Conventional vs. Islamic banks. Journal of Banking & Finance, 58, 418-435. Mollah, S., Hassan, M. K., Al Farooque, O., & Mobarek, A. (2017). The governance, risktaking, and performance of Islamic banks. Journal of Financial Services Research, 51(2), 195-219. Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial Economics, 5(2), 147-175. Pathan, S., & Faff, R. (2013). Does board structure in banks really affect their performance? Journal of Banking & Finance, 37(5), 1573-1589. Pathan, S., Wong, P. H., & Benson, K. (2019). How do ‘busy’and ‘overlap’directors relate to CEO pay structure and incentives?. Accounting & Finance, 59(2), 1341-1382. Penas, M. F., & Unal, H. (2004). Gains in bank mergers: Evidence from the bond markets. Journal of Financial Economics, 74(1), 149-179. Piot, C., & Missonier-Piera, F. (2009). Corporate governance reform and the cost of debt financing of listed French companies. Available at SSRN 960681. Rahaman, M. M., & Al Zaman, A. (2013). Management quality and the cost of debt: Does management matter to lenders? Journal of Banking & Finance, 37(3), 854-874. Reeb, D. M., Mansi, S. A., & Allee, J. M. (2001). Firm internationalization and the cost of debt financing: Evidence from non-provisional publicly traded debt. Journal of Financial and Quantitative Analysis, 36(3), 395-414. Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41-55. Safieddine, A. (2009). Islamic financial institutions and corporate governance: New Insights for Agency Theory. Corporate Governance: An International Review, 17: 142-158. Safiullah, M., & Shamsuddin, A. (2019). Risk-adjusted efficiency and corporate governance: Evidence from Islamic and conventional banks. Journal of Corporate Finance, 55, 105-140. Shivdasani, A. (1993). Board composition, ownership structure, and hostile takeovers. Journal of Accounting and Economics, 16(1-3), 167-198. Trinh, V. Q., Elnahass, M., Salama, A., & Izzeldin, M. (2019). Board busyness, performance and financial stability: does bank type matter? The European Journal of Finance, 1-28. Wintoki, M. B., Linck, J. S., & Netter, J. M. (2012). Endogeneity and the dynamics of internal corporate governance. Journal of Financial Economics, 105(3), 581-606. Wooldridge, J. M. (2009). On estimating firm-level production functions using proxy variables to control for unobservables. Economics Letters, 104(3), 112-114. World Bank (2016). “Labor market polarization in developing countries: challenges ahead.” Accessed 1 August 2017. http://blogs.worldbank.org. Yermack, D. (1996). Higher market valuation of companies with a small board of directors. Journal of Financial Economics, 40(2), 185-211.
24
Journal Pre-proof Table 1: Final Sample distributions for the Whole Sample Period Notes: The ultimate sample includes 70 listed banks (386 bank-year observations) with 27 IBs (150 bank-year observations) and 43 CBs (236 bank-year observations) for a six-year period from 2010. These banks are operating in 11 dual-banking countries. Country Observations Observations Observations % % % (IBs) (CBs) (Full Sample) (IBs) (CBs) (Full Sample) Bahrain 30 30 60 20.00 12.71 15.54 Bangladesh 36 44 80 24.00 18.64 20.73 Egypt 3 9 12 2.00 3.81 3.11 Indonesia 6 66 72 4.00 27.97 18.65 Jordan 12 29 41 8.00 12.29 10.62 Kuwait 3 12 15 2.00 5.09 3.89 Pakistan 24 6 30 16.00 2.54 7.77 Qatar 18 24 42 12.00 10.17 10.88 Saudi Arabia 6 6 12 4.00 2.54 3.11 UAE 6 6 12 4.00 2.54 3.11 Oman 6 4 10 4.00 1.70 2.59 Total bank-year observations 150 236 386 100 100 100 Number of banks 27 43 70 -
f o
l a
e
o r p
r P
n r u
o J
25
Journal Pre-proof Table 2: Variable Definitions and Sources Variables Abbreviations Cost of debt Cost of Debt
# Average directorships of directors % Busy directors Board of Directors Size Board Independence Chair-CEO duality Bank size Bank leverage GDP per capita Herfindahl-Hirschman Index Average country governance index Islamic banking dummy
Definitions After-tax weighted average cost of debt for the security. It is calculated by using government bond rates, a debt adjustment factor (DAF), the proportions of short-term and long-term debt to total debt, and the stock's effective tax rate. The debt adjustment factor represents the average yield above government bonds for a given rating class. Source: Bloomberg ABOD Average outside directorships per independent director, calculated as total number of additional (outside) boards held by independent directors divided by number of independent directors on the board. Source: annual report %BBOD Percentage of busy independent directors on the board (%), calculated as number of independent directors serving on two or more additional (outside) firms divided by number of independent directors on the board. Source: annual report LogBSIZE Natural logarithm of the total number of board of directors’ members. Source: annual report %IND Percentage of independent non-executive directors on the board of directors. Source: annual report DUAL Dummy variable which takes value of one if the chair and CEO is the same person and zero otherwise. Source: annual report LogTA Natural logarithm of the total assets. Source: DataStream LEV Total debt to total equity. Source: DataStream GDPCAPITA Natural logarithm of Gross domestic products per capita. Source: World Bank HHI The square of the sum of the ratio of total assets of each firm-year to total assets of all banks each year. Its range is from 0 to 1. Source: Bankscope GOV_COUNTRY The average value of six individual country governance measures including the corruption, government effectiveness, political stability, and regulatory quality, the rule of law, and voice and accountability. Source: World Bank ISLAMIC Dummy variable, taking value of one if the observed firm is classified as Islamic banks, otherwise zero. Source: Bankscope, annual report, and Central Bank’s website of each country.
f o
l a
e
o r p
r P
n r u
o J
26
Journal Pre-proof Table 3: Descriptive Statistics Notes: The table reports descriptive statistics of all variables employed in the main regression models of the study for the full sample. The ***, **, * represents p-values of 0.01, 0.05, and 0.10. See Table 2 and Section 6.1 for other variable definitions VARIABLES N Mean Median Std. Min Max Cost of Debt 386 2.281 1.394 2.532 0 10.320 ABOD 386 2.257 1.8 2.046 0 11 %BBOD 386 0.479 0.5 0.377 0 1 BSIZE 386 9.544 9 3.705 3 25 %IND 386 0.348 0.333333 0.237 0 1 DUAL 386 0.039 0 0.194 0 1 %INDQ 383 0.190 0 0.276 0 1 %INDEXP 386 0.294 0.2 0.339 0 1 LogASIZE 380 3.724 3 0.925 2 6 %BAC 380 0.495 0.5 0.333 0 1 INDMET 310 7.769 6 5.006 0 35 LogTA 386 15.408 15.427 1.287 11.999 18.047 LEV 386 8.039 7.775 3.774 -4.210 19.998 HHI 386 0.142 0.109 0.095 0.058 0.672 GDPCAPITA 386 8.750 8.216 1.542 6.634 11.480 GOV_COUNTRY 386 -0.286 -0.212 0.526 -1.181 0.737 ISLAMIC 386 0.389 0 0.488 0 1
f o
l a
e
o r p
r P
n r u
o J
27
Journal Pre-proof Table 4: Correlation Matrix Notes: The table presents the Pearson pair-wise correlation matrix significance at the 5% level. See Table 2 for variable definitions. (1) (2) (3) 1. ABOD 1 2. %BBOD 0.813* 1 3. LogBSIZE 0.128* 0.191* 1 4. %IND 0.165* 0.201* -0.403* 5. DUAL 0.078 0.064 0.060 6. LogTA -0.038 0.005 -0.039 7. LEV -0.080 -0.053 0.153* 8. GDPCAPITA 0.064 0.109* -0.153* 9. HHI 0.022 0.011 -0.024 10. GOV_COUNTRY 0.112* 0.127* -0.144* 11. ISLAMIC -0.072 -0.097 0.233*
among main independent variables used in our main analysis for full subsample. * indicates (4)
(5)
(6)
1 -0.035 0.218* -0.170* 0.260* -0.115* 0.323* -0.071
1 0.042 0.050 -0.070 0.019 -0.099 -0.160*
1 0.042 0.478* 0.071 0.430* -0.111*
l a
o J
28
f o
ro
-p
e r P
n r u
(7)
1 -0.449* -0.115* -0.489* 0.052
(8)
(9)
(10)
(11)
1 0.146* 0.630* -0.001
1 0.074 0.144*
1 -0.063
1
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 5: Cost of Debt and Board of Directors’ Busyness The table presents OLS regression results identifying the effect of busy boards of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt ABOD -0.353*** (0.000) ABOD*ISLAMIC 0.158** (0.045) %BBOD -1.623*** (0.000) %BBOD*ISLAMIC 0.763* (0.079) LogBSIZE -2.710*** -2.629*** (0.000) (0.000) %IND 1.685*** 1.802*** (0.001) (0.000) DUAL -1.985*** -2.111*** (0.000) (0.000) LogTA 0.587*** 0.598*** (0.000) (0.000) LEV -0.072** -0.071** (0.015) (0.015) GDPCAPITA -1.297*** -1.214*** (0.000) (0.000) HHI -3.471*** -3.601*** (0.000) (0.000) GOV_COUNTRY 1.910*** 1.673*** (0.000) (0.000) ISLAMIC -1.142*** -1.187*** (0.000) (0.000) Constant 13.000*** 11.826*** (0.000) (0.000) Year fixed effects YES YES Observations 386 386 R-squared 0.600 0.581
29
Journal Pre-proof Table 6a: Trend Analysis: Bank Cost of Debt by the Interval Categories of Board of Directors’ Busyness The table presents trend analysis results identifying a bank’s cost of debt for every incremental change in board busyness proxy measures. Panel A % busy outside directors Intervals <10% 10-19% 20-29% 30-39% 40-49% 50-59% 60-69% 70-79% 80-89% 90-99% N 117 1 15 17 8 62 41 38 7 0 % 30.3% 0.3% 3.9% 4.4% 2.1% 16.1% 10.6% 9.8% 1.8% 0 Cost of debt 2.842 0.659 0.313 0.096 0.089 0.010 1.916 2.790 3.267 Panel B Average number of outside directorships hold by outside directors Intervals <1 1-1.99 2-2.99 3-3.99 4-4.99 5-5.99 6-6.99 7-7.99 8-8.99 9-9.99 N 109 89 57 57 25 22 14 9 1 1 % 28.2% 23.1% 14.8% 14.8% 6.5% 5.7% 3.6% 2.3% 0.3% 0.3% Cost of debt 3.385 2.562 1.744 1.620 1.445 1.202 1.187 0.917 0.243 0.579
ro
f o
l a
p e
r P
n r u
o J
30
100% 80 20.7% 3.388 >9.99 2 0.4% 1.902
Journal Pre-proof Table 6b: Possible Non-linear Relationship between Cost of Debt and Board of Directors’ Busyness The table presents OLS regression results for the full sample (Panel A), conventional bank subsample (Panel B) and Islamic bank subsample (Panel C) identifying the possible nonlinear effect of busy board of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses.
VARIABLES ABOD (ABOD)2 %BBOD (%BBOD)2 LogBSIZE %IND DUAL LogTA LEV GDPCAPITA HHI GOV_COUNTRY ISLAMIC Constant Year fixed effects Observations
PANEL A: FULL SAMPLE (1) (2) -0.551*** (0.000) 0.035** (0.019) -2.402*** (0.010) 1.067 (0.217) -2.492*** -2.555*** (0.000) (0.000) 2.340*** 2.220*** (0.000) (0.000) -2.065*** -2.109*** (0.000) (0.000) 0.566*** 0.587*** (0.000) (0.000) -0.076*** -0.067** (0.010) (0.028) -1.192*** -1.183*** (0.000) (0.000) -3.662*** -3.782*** (0.000) (0.000) 1.625*** 1.577*** (0.000) (0.000) -0.840*** -0.840*** (0.000) (0.000) 11.800*** 11.361*** (0.000) (0.000) YES YES 386 386
PANEL C: ISLAMIC BANKS (5) (6) -0.028 (0.883) -0.006 (0.820) -0.031 (0.981) -0.326 (0.794) -2.338*** -2.297*** (0.000) (0.000) -0.341 -0.362 (0.522) (0.516) -
0.455*** (0.000) -0.011 (0.731) -0.869*** (0.003) -2.417** (0.012) 1.579** (0.036)
0.450*** (0.000) -0.008 (0.806) -0.840*** (0.002) -2.440*** (0.009) 1.493** (0.024)
12.437*** (0.000) YES 236
8.748*** (0.007) YES 150
8.450*** (0.006) YES 150
f o
l a
n r u
o J
PANEL B: CONVENTIONAL BANKS (3) (4) -0.768*** (0.000) 0.057*** (0.002) -4.312*** (0.000) 2.711** (0.019) -2.055*** -2.282*** (0.000) (0.000) 4.126*** 4.409*** (0.000) (0.000) -2.002*** -2.009*** (0.000) (0.000) 0.587*** 0.605*** (0.000) (0.000) -0.187*** -0.208*** (0.000) (0.000) -1.314*** -1.272*** (0.000) (0.000) -5.101*** -5.413*** (0.001) (0.000) 1.429*** 0.987* (0.005) (0.053)
o r p
e
r P
12.103*** (0.000) YES 236
31
Journal Pre-proof Adjusted R2
0.602
0.580
0.658
0.640
0.394
f o
l a
e
o r p
r P
n r u
o J
32
0.394
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 7: Cost of Debt and Board of Directors’ Busyness: The Effect of Agency Costs The table presents OLS regression results identifying the effect of agency costs on the relationship between busy boards of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt ABOD -0.350*** (0.000) ABOD*CASH -0.001 (0.996) ABOD*CASH* ISLAMIC 0.107*** (0.001) %BBOD -1.572*** (0.000) %BBOD*CASH 0.012 (0.961) %BBOD*CASH*ISLAMIC 0.765*** (0.000) LogBSIZE -2.607*** -2.556*** (0.000) (0.000) %IND 1.749*** 1.704*** (0.000) (0.000) DUAL -2.056*** -2.218*** (0.000) (0.000) LogTA 0.586*** 0.612*** (0.000) (0.000) LEV -0.047 -0.047* (0.101) (0.099) CASH -0.518*** -0.703*** (0.000) (0.000) GDPCAPITA -1.462*** -1.422*** (0.000) (0.000) HHI -2.664*** -2.521*** (0.000) (0.001) GOV_COUNTRY 2.559*** 2.462*** (0.000) (0.000) ISLAMIC -0.967*** -1.133*** (0.000) (0.000) Constant 14.551*** 13.814*** (0.000) (0.000) Year fixed effects YES YES Observations 386 386 R-squared 0.625 0.608
33
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 8: Sensitivity Test: Adding More Corporate Governance Variables The table presents OLS regression results identifying the effect of busy board of directors and other corporate governance variables on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 and Section 6.1 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt ABOD -0.374*** (0.000) ABOD*ISLAMIC 0.300*** (0.001) %BBOD -2.110*** (0.000) %BBOD*ISLAMIC 1.975*** (0.000) %INDQ 0.339 0.310 (0.360) (0.416) %INDEXP 0.298 0.182 (0.310) (0.557) LogASIZE 0.109 0.199 (0.440) (0.139) %BAC -0.850** -0.701** (0.011) (0.036) INDMET -0.012 -0.008 (0.601) (0.729) LogBSIZE -2.957*** -2.873*** (0.000) (0.000) %IND 0.540 0.618 (0.335) (0.291) DUAL -0.984** -1.226*** (0.036) (0.000) LogTA 0.698*** 0.666*** (0.000) (0.000) LEV -0.060* -0.061* (0.078) (0.071) GDPCAPITA -1.007*** -0.879*** (0.000) (0.000) HHI -3.907*** -3.868*** (0.000) (0.000) GOV_COUNTRY 1.321*** 1.084*** (0.002) (0.008) ISLAMIC -1.662*** -1.954*** (0.000) (0.000) Constant 9.663*** 8.566*** (0.000) (0.000) Year fixed effects YES YES Observations 310 310 R-squared 0.653 0.641
34
Journal Pre-proof Table 9: Alternative measures for Cost of Debt The table presents OLS regression results using alternative measures for cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust Pvalues in parentheses. See Table 2 for variable definitions. Panel A: Conventional Banks Panel B: Islamic Banks (1) (2) (3) (4) VARIABLES Interest Interest Return Return -0.355*** 0.768 ABOD (0.002)
(0.111)
5.067** (0.031)
0.803 (0.453) 3.452* (0.052)
Constant
YES 236 0.321
YES 236 0.344
YES 150 0.123
YES 150 0.156
Jo ur
na
lP
re
-p
ro
Year fixed effects Observations R-squared
of
1.432*** (0.000)
-0.541** (0.040) 0.843*** (0.000)
%BBOD
35
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Table 10: Endogeneity Treatment: Using one-year lag of independent variables The table presents OLS regression results identifying the effect of busy board of directors on a bank’s cost of debt. Robust standard errors are adjusted for heteroscedasticity. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. (1) (2) VARIABLES Cost of Debt Cost of Debt L.ABOD -0.391*** (0.000) L.ABOD*ISLAMIC 0.202** (0.027) L.%BBOD -1.979*** (0.000) L.%BBOD*ISLAMIC 1.212** (0.019) L.LogBSIZE -2.810*** -2.645*** (0.000) (0.000) L.%IND 1.673*** 1.809*** (0.001) (0.001) L.DUAL -2.118*** -2.249*** (0.000) (0.000) L.LogTA 0.616*** 0.624*** (0.000) (0.000) L.LEV -0.077** -0.081** (0.016) (0.011) L.GDPCAPITA -1.351*** -1.264*** (0.000) (0.000) L.HHI -3.134*** -2.989*** (0.000) (0.000) L.GOV_COUNTRY 1.992*** 1.780*** (0.000) (0.000) ISLAMIC -1.194*** -1.376*** (0.000) (0.000) Constant 12.555*** 11.293*** (0.000) (0.000) Year fixed effects YES YES Observations 316 316 R-squared 0.603 0.588
36
Journal Pre-proof Table 11: Robustness Check: Endogeneity Treatment by Two-stage Least Square, Three-stage Least Square and GMM The table presents 2SLS and 3SLS and GMM regression results identifying the effect of busy board of directors and other corporate governance variables on a bank’s cost of debt. *** p<0.01, ** p<0.05, * p<0.1. Robust P-values in parentheses. See Table 2 for variable definitions. Panel A: 2SLS Panel B: 3SLS Panel C: GMM VARIABLES (1) (2) (3) (4) (5) (6) Cost of Debt Cost of Debt Cost of Debt Cost of Debt Cost of Debt Cost of Debt ABOD -1.910*** -2.061*** -0.225*** (0.005) (0.000) (0.002) ABOD*ISLAMIC 1.711*** 1.865*** 0.195** (0.009) (0.000) (0.038) %BBOD -13.444*** -14.803*** -1.520*** (0.002) (0.000) (0.001) %BBOD*ISLAMIC 11.782*** 13.016*** 1.454** (0.003) (0.001) (0.041) LogBSIZE -1.090 0.551 -0.914 0.952 -0.946** -1.031 (0.271) (0.656) (0.178) (0.437) (0.017) (0.224) %IND 2.073** 4.055*** 2.095*** 4.319*** -0.497 0.533 (0.025) (0.004) (0.009) (0.001) (0.442) (0.593) DUAL -0.941 -1.197 -0.819 -1.059 -0.373 0.183 (0.476) (0.331) (0.336) (0.306) (0.336) (0.888) LogTA 0.345** 0.253 0.347** 0.248 0.275* 0.294 (0.021) (0.206) (0.033) (0.253) (0.065) (0.257) LEV -0.048 -0.082* -0.043 -0.080 -0.021 -0.027 (0.204) (0.096) (0.378) (0.190) (0.561) (0.658) GDPCAPITA -1.675*** -1.237*** -1.678*** -1.188*** -0.477*** -0.394* (0.000) (0.000) (0.000) (0.000) (0.004) (0.093) HHI 0.774 2.114 0.486 1.678 -1.582*** -1.196** (0.732) (0.415) (0.823) (0.566) (0.000) (0.017) GOV_COUNTRY 3.638*** 2.734*** 3.721*** 2.729*** 0.988** 0.700 (0.000) (0.004) (0.000) (0.006) (0.012) (0.213) ISLAMIC -5.119*** -7.458*** -5.495*** -8.140*** -0.568* -0.867** (0.002) (0.001) (0.000) (0.000) (0.074) (0.033) Constant 19.630*** 15.042*** 20.066*** 15.056*** 3.347* 3.719
f o
l a
e
o r p
r P
n r u
o J
37
Journal Pre-proof (0.000)
(0.000)
(0.000)
(0.000)
YES 2SLS 386 0.000 0.000 0.009
YES 2SLS 386 0.000 0.000 0.100
YES 3SLS 386 0.000
YES 3SLS 386 0.000
0.122 0.000
0.344 0.000
Cost of Debtt-1 Year fixed effects Model Observations Wald 2 (p-value) Endogeneity (p-value) Over identification (p-value) Hansen-Sargan test (p-value) Breusch-Pagan LM test (p-value) AR(1) p-value AR(2) p-value Hansen test (p-value)
l a
e
o r p
r P
n r u
o J
38
f o
(0.071) 0.786*** (0.000) YES GMM 316 0.000
(0.238) 0.766*** (0.000) YES GMM 316 0.000
0.000 0.440 0.202
0.000 0.320 0.328
Journal Pre-proof Table 12: Propensity score matching method: The effect of board busyness on the cost of debt (Full sample) Estimator: propensity-score matching Treatment model: probit Dependent variable: cost of debt Panel A: Average treatment effects with nearest neighbour Matching method Treated Control Δ S.E. 1:1 matching without replacement Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.954 3.211 -1.256*** 0.304 1:1 matching with replacement Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.679 3.245 -1.566*** 0.502 Nearest neighbour (n=2) Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.679 2.847 -1.168** 0.414 Nearest neighbour (n=3) Unmatched 1.671 3.211 -1.539*** 0.252 Matched 1.679 2.935 -1.257*** 0.399 Panel B: Average treatment effect on the treated with 1:1 nearest neighbour matching and bootstrapping of standard errors No of obs. Replications Observed (Δ) Bias S.E. 388 100 -1.548*** 0.185 0.369 388 1000 -1.548*** 0.209 0.342 388 10000 -1.548*** 0.211 0.342 Panel C: Regressions on matched samples (1) (2) (3) (4) Independent variables 1:1 matching 1:1 matching Nearest Nearest neighbour without replacement with replacement neighbour (n=3) (n=2) BOD Busyness Dummy -1.221*** -1.197*** -1.014*** -0.999*** (0.000) (0.000) (0.000) (0.000) Controls Yes Yes Yes Yes Constant 12.817*** 16.590*** 14.421*** 13.891***
f o
l a
e
o r p
r P
n r u
o J
39
T-stat -6.11 -4.14 -6.11 -3.12 -6.11 -2.82 -6.11 -3.15 T-stat -4.19 -4.53 -4.53
Journal Pre-proof (0.000) (0.000) (0.000) (0.000) Adjusted R-squared 0.588 0.579 0.520 0.541 Observations 306 464 344 356 Note: The table present the propensity score matching technique showing the results of the average treatment effects (ATE) and the average treatment effect on the treated (ATT) with 1:1 nearest neighbour matching and bootstrapping of standard errors. The ATE and ATT of board busyness on the cost of debt (Δ) is estimated as the difference between the mean changes of banks having board with busy outside directors (column “Treated”) and that of matched banks having board without busy outside directors (column “Non-treated”). P-value is presented in parentheses. Tstatistics based on standard errors are presented in final column. *** p<0.01, ** p<0.05, * p<0.1.
f o
l a
e
o r p
r P
n r u
o J
40
Journal Pre-proof
Jo ur
na
lP
re
-p
ro
of
Appendix 1: Unmatched versus Matched sample: propensity score
41
Journal Pre-proof
Highlights The study documents that busy boards of directors enhance
Jo ur
na
lP
re
-p
ro
of
a bank’s financing capacity by lowering its cost of debt. Such negative association is more pronounced in conventional banks than their Islamic counterparts. Results survive across alternative proxies for cost of debt and board busyness, and alternative model specifications and methodology. Our results contribute to understanding the indispensable role busy boards play in debt financing
42